Outros artigos de André Saddy Artigo doutrinário
Adoption of disruptive technologies in brazilian public administration: diagnosis and prospects
Citação acadêmica
Copie a referência deste artigo no estilo da sua escolha e cole no seu trabalho. Geramos a citação no padrão exigido pela maioria das revistas jurídicas brasileiras. Ver guia completo.
Ver prévia das três referências▸
ABNT
SADDY, André. Adoption of disruptive technologies in brazilian public administration: diagnosis and prospects. professor_upload, 2026. Acesso via: JurisTube — Acervo Digital de Direito. Disponível em: https://juristube.com.br/colunistas/andre-saddy/adoption-of-disruptive-technologies-in-brazilian-public-administration-diagnosis. Acesso em: 24 set. 2026.
APA
Saddy, A. (2026). Adoption of disruptive technologies in brazilian public administration: diagnosis and prospects. *professor_upload*. https://juristube.com.br/colunistas/andre-saddy/adoption-of-disruptive-technologies-in-brazilian-public-administration-diagnosis
Chicago
SADDY, André. 2026. “Adoption of disruptive technologies in brazilian public administration: diagnosis and prospects.” *professor_upload*. https://juristube.com.br/colunistas/andre-saddy/adoption-of-disruptive-technologies-in-brazilian-public-administration-diagnosis
BibTeX
@article{andr-saddy-adoption-of-disruptive-technologies-in-b-2026,
author = {Saddy, André},
title = {Adoption of disruptive technologies in brazilian public administration: diagnosis and prospects},
journal = {professor_upload},
year = {2026},
url = {https://juristube.com.br/colunistas/andre-saddy/adoption-of-disruptive-technologies-in-brazilian-public-administration-diagnosis}
}_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
1
ADOPTION OF DISRUPTIVE TECHNOLOGIES IN
BRAZILIAN PUBLIC ADMINISTRATION: DIAGNOSIS
AND PROSPECTS
Andrè SADDY1, Raphael Lobato Collet Janny TEIXEIRA2
_____________________________________________________________
INDEX
1. INTRODUCTION
1 Post-Doctorate from the Centre for Socio-Legal Studies of the Faculty of Law of the University of Oxford,
European Doctorate in "Current Problems of Administrative Law" from the Faculty of Law of the Complutense
University of Madrid, with the support of Becas Complutense Predoctorales en España, Master's degree in Public
Administration from the Faculty of Law of the University of Lisbon, with support from the Alßan Programme, the
European Union's High Level Scholarship Programme for Latin America, postgraduate degree in Public Regulation
and Competition from the Centre for Studies in Public Law and Regulation (CEDIPRE) of the Faculty of Law of
the University of Coimbra. Professor of administrative law at the Faculty of Law, Masters in Constitutional Law
and Doctorate in Rights, Institutions and Business at the Fluminense Federal University (UFF). Professor of
administrative law at the Law Department of the Pontifical Catholic University of Rio de Janeiro (PUC-Rio).
Director of Standardisation at the Brazilian Institute of Administrative Law (IBDA). Vice-President of the Rio de
Janeiro Administrative Law Institute (IDARJ). President of the Centre for Empirical-Legal Studies (CEEJ). Founder
and Coordinator of the Contemporary Administrative Law Research, Teaching and Extension Group (GDAC).
Founding partner of Saddy Advogados. Consultant and legal expert.
2 PhD student in Law at the State University of Rio de Janeiro (UERJ). Master's in Constitutional Law from
Fluminense Federal University (Uff). Postgraduate (latu sensu) in industrial property from Rio de Janeiro State
University (UERJ). MBA in management from the Dom Cabral Foundation (FDC). Master Lawyer at Petróleo
Brasileiro S/A (Petrobras), specialising in Digital Law, Data Protection and New Technologies. Coordinator of the
postgraduate programme in Digital Law at Faculdade Mar Atlântico (FMA). Executive Director of Research at the
Centre for Empirical-Legal Studies (CEEJ).
______________________________________________________________________________
Copyleft – Ius Publicum
1
ADOPTION OF DISRUPTIVE TECHNOLOGIES IN
BRAZILIAN PUBLIC ADMINISTRATION: DIAGNOSIS
AND PROSPECTS
Andrè SADDY1, Raphael Lobato Collet Janny TEIXEIRA2
_____________________________________________________________
INDEX
1. INTRODUCTION
1 Post-Doctorate from the Centre for Socio-Legal Studies of the Faculty of Law of the University of Oxford,
European Doctorate in "Current Problems of Administrative Law" from the Faculty of Law of the Complutense
University of Madrid, with the support of Becas Complutense Predoctorales en España, Master's degree in Public
Administration from the Faculty of Law of the University of Lisbon, with support from the Alßan Programme, the
European Union's High Level Scholarship Programme for Latin America, postgraduate degree in Public Regulation
and Competition from the Centre for Studies in Public Law and Regulation (CEDIPRE) of the Faculty of Law of
the University of Coimbra. Professor of administrative law at the Faculty of Law, Masters in Constitutional Law
and Doctorate in Rights, Institutions and Business at the Fluminense Federal University (UFF). Professor of
administrative law at the Law Department of the Pontifical Catholic University of Rio de Janeiro (PUC-Rio).
Director of Standardisation at the Brazilian Institute of Administrative Law (IBDA). Vice-President of the Rio de
Janeiro Administrative Law Institute (IDARJ). President of the Centre for Empirical-Legal Studies (CEEJ). Founder
and Coordinator of the Contemporary Administrative Law Research, Teaching and Extension Group (GDAC).
Founding partner of Saddy Advogados. Consultant and legal expert.
2 PhD student in Law at the State University of Rio de Janeiro (UERJ). Master's in Constitutional Law from
Fluminense Federal University (Uff). Postgraduate (latu sensu) in industrial property from Rio de Janeiro State
University (UERJ). MBA in management from the Dom Cabral Foundation (FDC). Master Lawyer at Petróleo
Brasileiro S/A (Petrobras), specialising in Digital Law, Data Protection and New Technologies. Coordinator of the
postgraduate programme in Digital Law at Faculdade Mar Atlântico (FMA). Executive Director of Research at the
Centre for Empirical-Legal Studies (CEEJ).
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
2
2. NEW DISRUPTIVE TECHNOLOGIES AND PUBLIC ADMISTRATION
2.1 Public economic promotion of technological innovation
2.2 Infrastructure law and new disruptive technologies
2.3 Use of Blockchain by the Public Administration
2.4 IoT and Public Administration
3. AI AND PUBLIC ADMINISTRATION
3.1 Public policies and investment plans in AI
3.2 Regulation and regulation of AI in Public Administration
3.3 Administrative decision-making using AI
4. CONCLUSIONS
1. INTRODUCTION
Disruptive innovations can be understood as substantial, unpredictable changes that
occur irregularly, creating new products, manufacturing processes or business models. In
addition, disruptive innovations generally reduce or even hijack the market share of
established companies or create new markets. As argued in another paper3, “Disruptive
innovations disrupt, break and fracture the course of the hitherto existing economic process,
[...] whether for the provision of a service or the supply of a good”.
The adoption of new technologies in public administration is currently seen as an
indispensable process for modernising government services and improving interaction with
citizens. In January 2024, the Organisation for Economic Co-operation and Development
3 A. SADDY, Perspectivas do direito da infraestrutura com o surgimento das novas tecnologias (inovações)
disruptivas, in A. SADDY et al., Aspectos Jurídicos das Novas Tecnologias (inovações) Disruptivas, Rio de Janeiro,
CEEJ, 2 ed., 2020, 61-62.
______________________________________________________________________________
Copyleft – Ius Publicum
2
2. NEW DISRUPTIVE TECHNOLOGIES AND PUBLIC ADMISTRATION
2.1 Public economic promotion of technological innovation
2.2 Infrastructure law and new disruptive technologies
2.3 Use of Blockchain by the Public Administration
2.4 IoT and Public Administration
3. AI AND PUBLIC ADMINISTRATION
3.1 Public policies and investment plans in AI
3.2 Regulation and regulation of AI in Public Administration
3.3 Administrative decision-making using AI
4. CONCLUSIONS
1. INTRODUCTION
Disruptive innovations can be understood as substantial, unpredictable changes that
occur irregularly, creating new products, manufacturing processes or business models. In
addition, disruptive innovations generally reduce or even hijack the market share of
established companies or create new markets. As argued in another paper3, “Disruptive
innovations disrupt, break and fracture the course of the hitherto existing economic process,
[...] whether for the provision of a service or the supply of a good”.
The adoption of new technologies in public administration is currently seen as an
indispensable process for modernising government services and improving interaction with
citizens. In January 2024, the Organisation for Economic Co-operation and Development
3 A. SADDY, Perspectivas do direito da infraestrutura com o surgimento das novas tecnologias (inovações)
disruptivas, in A. SADDY et al., Aspectos Jurídicos das Novas Tecnologias (inovações) Disruptivas, Rio de Janeiro,
CEEJ, 2 ed., 2020, 61-62.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
3
(OECD) published a study on the maturity of digital government4. According to this
diagnosis, the digital transformation in Brazilian public administration has progressed at a
moderate pace, performing close to the average of the OECD countries5.
Another study, the ICT Electronic Government Survey 20236, carried out by the
Brazilian Internet Steering Committee revealed significant advances in the adoption of
emerging technologies by federal and state public bodies in the Executive, Legislative and
Judiciary branches. Artificial Intelligence (AI) was the highlight among disruptive
technologies, with growth from 24 per cent in 2021 to 30 per cent in 2023, being widely used
by the Judiciary (68 per cent of agencies). In addition, the Internet of Things (IoT) saw an
increase in adoption, from 18 per cent to 27 per cent, while the use of Blockchain remained
stable, varying from 13 per cent to 15 per cent. Cloud services were also widely implemented,
especially for corporate emails (81 per cent of federal agencies and 59 per cent of state
agencies).
This data reflects the progress of digital transformation in the public sector, with a
focus on automation, connectivity and greater interaction with society, although some
technologies, such as Blockchain, still face challenges for wider adoption. In fact, Brazil has
a relevant market and a society that is open to adopting innovations, despite not being a major
4 The following points were taken into account in the evaluation: i) Digital by Design: The government is planned
to be digital from the outset, with technologies integrated into its processes. ii) Data-driven Public Sector: Decisions
and improvements in public services are based on the use of data. iii) Government as a Platform: The government
provides platforms and resources for use by various sectors. iv) Open by Default: Transparency and access to
information. v) User-driven: Services and policies prioritise users' needs and experiences. vi) Proactiveness: The
government anticipates citizens' needs and offers solutions in advance. See OCDE 2023 OECD Digital Government
Index.
5 Ibid.
6 Núcleo de informação e coordenação do ponto br. Pesquisa sobre o uso das Tecnologias de Informação e
Comunicação no setor público brasileiro - TIC Governo Eletrônico 2023.
______________________________________________________________________________
Copyleft – Ius Publicum
3
(OECD) published a study on the maturity of digital government4. According to this
diagnosis, the digital transformation in Brazilian public administration has progressed at a
moderate pace, performing close to the average of the OECD countries5.
Another study, the ICT Electronic Government Survey 20236, carried out by the
Brazilian Internet Steering Committee revealed significant advances in the adoption of
emerging technologies by federal and state public bodies in the Executive, Legislative and
Judiciary branches. Artificial Intelligence (AI) was the highlight among disruptive
technologies, with growth from 24 per cent in 2021 to 30 per cent in 2023, being widely used
by the Judiciary (68 per cent of agencies). In addition, the Internet of Things (IoT) saw an
increase in adoption, from 18 per cent to 27 per cent, while the use of Blockchain remained
stable, varying from 13 per cent to 15 per cent. Cloud services were also widely implemented,
especially for corporate emails (81 per cent of federal agencies and 59 per cent of state
agencies).
This data reflects the progress of digital transformation in the public sector, with a
focus on automation, connectivity and greater interaction with society, although some
technologies, such as Blockchain, still face challenges for wider adoption. In fact, Brazil has
a relevant market and a society that is open to adopting innovations, despite not being a major
4 The following points were taken into account in the evaluation: i) Digital by Design: The government is planned
to be digital from the outset, with technologies integrated into its processes. ii) Data-driven Public Sector: Decisions
and improvements in public services are based on the use of data. iii) Government as a Platform: The government
provides platforms and resources for use by various sectors. iv) Open by Default: Transparency and access to
information. v) User-driven: Services and policies prioritise users' needs and experiences. vi) Proactiveness: The
government anticipates citizens' needs and offers solutions in advance. See OCDE 2023 OECD Digital Government
Index.
5 Ibid.
6 Núcleo de informação e coordenação do ponto br. Pesquisa sobre o uso das Tecnologias de Informação e
Comunicação no setor público brasileiro - TIC Governo Eletrônico 2023.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
4
research and development centre. This is due, among other reasons, to the fact that the
availability of private capital is lower than in developed countries, requiring greater state
intervention to stimulate investment in new technologies.
However, in order to remain competitive and make significant progress, the country
needs to invest more in emerging technologies such as AI and overcome structural
challenges. Among these challenges7 , we highlight the inequalities in access to technologies,
which vary significantly between the different levels of government, regions of the country
and age groups of the population, highlighting the urgency of public policies that promote
digital inclusion and equity.
Against this backdrop, this article aims to present an overview of the authors'
concerns about new disruptive technologies, AI and public administration. To this end, the
article will be divided into two main topics.
The first topic will cover aspects relating to new disruptive technologies in general,
with a focus on public economic promotion of technological innovation; repercussions on
infrastructure law and the adoption of Blockchain and IoT in Public Administration.
The second topic deals with aspects relating to AI, in view of the special prominence
that this technology has been gaining. In this way, a diagnosis of the adoption of AI by the
Public Administration will be presented, with a special focus on the Brazilian Strategy for
7 The ICT eGovernment Survey (2023) revealed important challenges faced by Brazil in implementing digital
solutions in the public sector. Among the main obstacles are limitations in technical capacity and organisational
adaptation, as well as inequalities in the provision of digital services between the different levels of government and
regions, with a significant impact on smaller municipalities. There is also a prioritisation of technologies aimed at
tax collection and the internal activities of administrations, to the detriment of the development of tools aimed
directly at citizens. This technological imbalance particularly affects vulnerable populations, who face access
difficulties made worse by the reduction in public service centres. In addition, the growth of monitoring centres,
often without adequate regulation, raises concerns about privacy and data protection, highlighting the need for public
policies that promote an inclusive and responsible digital transformation.
______________________________________________________________________________
Copyleft – Ius Publicum
4
research and development centre. This is due, among other reasons, to the fact that the
availability of private capital is lower than in developed countries, requiring greater state
intervention to stimulate investment in new technologies.
However, in order to remain competitive and make significant progress, the country
needs to invest more in emerging technologies such as AI and overcome structural
challenges. Among these challenges7 , we highlight the inequalities in access to technologies,
which vary significantly between the different levels of government, regions of the country
and age groups of the population, highlighting the urgency of public policies that promote
digital inclusion and equity.
Against this backdrop, this article aims to present an overview of the authors'
concerns about new disruptive technologies, AI and public administration. To this end, the
article will be divided into two main topics.
The first topic will cover aspects relating to new disruptive technologies in general,
with a focus on public economic promotion of technological innovation; repercussions on
infrastructure law and the adoption of Blockchain and IoT in Public Administration.
The second topic deals with aspects relating to AI, in view of the special prominence
that this technology has been gaining. In this way, a diagnosis of the adoption of AI by the
Public Administration will be presented, with a special focus on the Brazilian Strategy for
7 The ICT eGovernment Survey (2023) revealed important challenges faced by Brazil in implementing digital
solutions in the public sector. Among the main obstacles are limitations in technical capacity and organisational
adaptation, as well as inequalities in the provision of digital services between the different levels of government and
regions, with a significant impact on smaller municipalities. There is also a prioritisation of technologies aimed at
tax collection and the internal activities of administrations, to the detriment of the development of tools aimed
directly at citizens. This technological imbalance particularly affects vulnerable populations, who face access
difficulties made worse by the reduction in public service centres. In addition, the growth of monitoring centres,
often without adequate regulation, raises concerns about privacy and data protection, highlighting the need for public
policies that promote an inclusive and responsible digital transformation.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
5
Artificial Intelligence (EBIA) and the AI Plan for the Good of All (PBIA). In addition, issues
relating to the regulation and regulation of AI in public administration will be addressed and,
finally, legal aspects relating to administrative decision-making using AI will be analysed.
In terms of methodology, bibliographical research was carried out, using doctrinal
and legal sources, especially scientific articles, periodicals, books and legislation on the
subject.
2. NEW DISRUPTIVE TECHNOLOGIES AND PUBLIC
ADMINISTRATION
2.1 Public economic support for technological innovation
Rodrigo da Fonseca Chauvet8, when dealing with public economic promotion of
technological innovation, understood here as the adoption by the Public Administration of
stimulating, inciting and inducing activities of an economic nature in the private sector, points
out that in the case of developing countries such as Brazil, the state has a fundamental role
in making direct investments in innovation development (RD&I).
This is because research, development and innovation activities generally require a
large amount of long-term investment and no guarantee of success, since technological risk
is inherent in the development of disruptive solutions. Consequently, the need for the state to
8 R.F. CHAUVET, Fomento público econômico à inovação tecnológica, in A. SADDY et al., Aspectos Jurídicos das
Novas Tecnologias (inovações) Disruptivas, cit., 92-123.
______________________________________________________________________________
Copyleft – Ius Publicum
5
Artificial Intelligence (EBIA) and the AI Plan for the Good of All (PBIA). In addition, issues
relating to the regulation and regulation of AI in public administration will be addressed and,
finally, legal aspects relating to administrative decision-making using AI will be analysed.
In terms of methodology, bibliographical research was carried out, using doctrinal
and legal sources, especially scientific articles, periodicals, books and legislation on the
subject.
2. NEW DISRUPTIVE TECHNOLOGIES AND PUBLIC
ADMINISTRATION
2.1 Public economic support for technological innovation
Rodrigo da Fonseca Chauvet8, when dealing with public economic promotion of
technological innovation, understood here as the adoption by the Public Administration of
stimulating, inciting and inducing activities of an economic nature in the private sector, points
out that in the case of developing countries such as Brazil, the state has a fundamental role
in making direct investments in innovation development (RD&I).
This is because research, development and innovation activities generally require a
large amount of long-term investment and no guarantee of success, since technological risk
is inherent in the development of disruptive solutions. Consequently, the need for the state to
8 R.F. CHAUVET, Fomento público econômico à inovação tecnológica, in A. SADDY et al., Aspectos Jurídicos das
Novas Tecnologias (inovações) Disruptivas, cit., 92-123.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
6
contribute resources to technological development will be greater, acting in a subsidiary way
to private initiative9.
In order to stimulate innovation and scientific and technological development in
Brazil, Constitutional Amendment 85/2015 introduced into Article 218 of the Constitution
the obligation for the state to promote and encourage scientific development, research,
scientific and technological training and innovation. It is worth noting that, until then, there
had been no mention of "innovation" in the constitutional text.
With the advent of Law No. 13,243/2016, identified as the Science, Technology and
Innovation Framework, important changes were made to legislation on the subject,
representing an important step towards stimulating scientific and technological development
in the country.
Despite the constitutional and legal provisions relating to public support for
innovation and technological development, Elisa Mara Coimbra10 points out that in Brazil,
political instability and a lack of resources end up affecting public support for innovation.
Thus, despite the fact that art. 10 of Law 11.540/2007 establishes a series of revenues, such
as the percentage of royalties on oil or gas production, to increase the resources of the
National Fund for Scientific and Technological Development (FNDCT), the federal
government has contingenced these resources to meet fiscal surplus targets, jeopardising the
planning of strategic actions in science and technology. The contingency of the FNDCT is
particularly damaging, as it affects not only the instruments directly, but also the equalisation
of funding needed to make interest rates lower than those of the market.
9 Ibid., 102.
10 E.M. COIMBRA, Desafios do fomento público à inovação: perspectivas e possibilidades, in A. SADDY et al.,
Aspectos Jurídicos das Novas Tecnologias (inovações) Disruptivas, cit., 139-142.
______________________________________________________________________________
Copyleft – Ius Publicum
6
contribute resources to technological development will be greater, acting in a subsidiary way
to private initiative9.
In order to stimulate innovation and scientific and technological development in
Brazil, Constitutional Amendment 85/2015 introduced into Article 218 of the Constitution
the obligation for the state to promote and encourage scientific development, research,
scientific and technological training and innovation. It is worth noting that, until then, there
had been no mention of "innovation" in the constitutional text.
With the advent of Law No. 13,243/2016, identified as the Science, Technology and
Innovation Framework, important changes were made to legislation on the subject,
representing an important step towards stimulating scientific and technological development
in the country.
Despite the constitutional and legal provisions relating to public support for
innovation and technological development, Elisa Mara Coimbra10 points out that in Brazil,
political instability and a lack of resources end up affecting public support for innovation.
Thus, despite the fact that art. 10 of Law 11.540/2007 establishes a series of revenues, such
as the percentage of royalties on oil or gas production, to increase the resources of the
National Fund for Scientific and Technological Development (FNDCT), the federal
government has contingenced these resources to meet fiscal surplus targets, jeopardising the
planning of strategic actions in science and technology. The contingency of the FNDCT is
particularly damaging, as it affects not only the instruments directly, but also the equalisation
of funding needed to make interest rates lower than those of the market.
9 Ibid., 102.
10 E.M. COIMBRA, Desafios do fomento público à inovação: perspectivas e possibilidades, in A. SADDY et al.,
Aspectos Jurídicos das Novas Tecnologias (inovações) Disruptivas, cit., 139-142.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
7
One solution to mitigate this problem would be to transform the nature of the
FNDCT from accounting to financial, transferring the resources from the National Treasury's
Single Account to an official financial institution, such as Finep. It's important to note that
Finep currently manages resources from various sectoral funds with specific purposes, but
they all suffer from problems related to contingency11.
2.2 Infrastructure law and new disruptive technologies
Many of the most recent disruptive innovations affect existing public infrastructures
and infrastructure law, requiring infrastructural changes for subsistence, either for their
evolution or for their implementation, forcing the market to behave differently from the way
it had been behaving, forcing public authorities and private initiative to expand and,
sometimes, to extinguish certain infrastructures. In this regard, we would emphasise what has
already been said in another text12:
These new disruptive technologies (innovations) have revolutionised the world we
live in and often require investment in infrastructure not by those who undertake these niches,
but by the public authorities, so that the sharing, collaborative, green, creative, multi-currency
economy and other existing technologies can be used to their full potential. If, historically,
infrastructures were built by those who would later exploit the activities, or by society as a
whole, today these economies point to a reality in which their creators use other people's
infrastructure, without questioning whether or not the infrastructure owner needs to grant
access to their networks because this is an essential facility.
[...]
11 Ibid, 142.
12 A. SADDY, Perspectivas do direito da infraestrutura com o surgimento das novas tecnologias (inovações)
disruptivas. In A. SADDY et al., Aspectos Jurídicos das Novas Tecnologias (inovações) Disruptivas, cit., 64-67.
______________________________________________________________________________
Copyleft – Ius Publicum
7
One solution to mitigate this problem would be to transform the nature of the
FNDCT from accounting to financial, transferring the resources from the National Treasury's
Single Account to an official financial institution, such as Finep. It's important to note that
Finep currently manages resources from various sectoral funds with specific purposes, but
they all suffer from problems related to contingency11.
2.2 Infrastructure law and new disruptive technologies
Many of the most recent disruptive innovations affect existing public infrastructures
and infrastructure law, requiring infrastructural changes for subsistence, either for their
evolution or for their implementation, forcing the market to behave differently from the way
it had been behaving, forcing public authorities and private initiative to expand and,
sometimes, to extinguish certain infrastructures. In this regard, we would emphasise what has
already been said in another text12:
These new disruptive technologies (innovations) have revolutionised the world we
live in and often require investment in infrastructure not by those who undertake these niches,
but by the public authorities, so that the sharing, collaborative, green, creative, multi-currency
economy and other existing technologies can be used to their full potential. If, historically,
infrastructures were built by those who would later exploit the activities, or by society as a
whole, today these economies point to a reality in which their creators use other people's
infrastructure, without questioning whether or not the infrastructure owner needs to grant
access to their networks because this is an essential facility.
[...]
11 Ibid, 142.
12 A. SADDY, Perspectivas do direito da infraestrutura com o surgimento das novas tecnologias (inovações)
disruptivas. In A. SADDY et al., Aspectos Jurídicos das Novas Tecnologias (inovações) Disruptivas, cit., 64-67.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
8
It can therefore be deduced from everything seen so far that the sharing,
collaborative, green, creative and multi-currency economies, as well as other technologies,
may not develop their own infrastructures. Their creators and administrators, as a rule, use
existing public infrastructure to provide their services or dispose of their goods. Furthermore,
there are situations in which the necessary infrastructure for these new technologies does not
exist and, therefore, their creation will depend on the public authorities for the development
of disruptive innovation.
To illustrate this situation13, has already cited the case of platforms known as Over
The Top (OTT), which use telecommunications infrastructure to offer services that often end
up competing with the infrastructure operators themselves. This situation occurred when, in
2015, the WhatsApp application began offering voice calls without its own transmission
infrastructure. Mobile operators claimed that this practice jeopardised their very existence,
damaging the infrastructure needed for the telecommunications sector.
As is well known, public infrastructure is crucial to the country's economic recovery
and progress. However, investments in public infrastructure are complex, and will become
even more difficult with the emergence of disruptive innovations. Changes in the costs and
externalities of network infrastructures will be inevitable, since financing and investment will
no longer be carried out exclusively by the public authorities and other contributions will be
necessary, such as charges for those who use the services.
In view of this, we have already had the opportunity to conclude14 that the study of
infrastructure law, which has traditionally been limited to aspects related to fixed, common
and sunk costs, the idea of network externalities, the concept of public networks and natural
monopolies, among other topics, will be affected by these new economies. As a result, some
13 Ibid., 65.
14 Ibid., 69.
______________________________________________________________________________
Copyleft – Ius Publicum
8
It can therefore be deduced from everything seen so far that the sharing,
collaborative, green, creative and multi-currency economies, as well as other technologies,
may not develop their own infrastructures. Their creators and administrators, as a rule, use
existing public infrastructure to provide their services or dispose of their goods. Furthermore,
there are situations in which the necessary infrastructure for these new technologies does not
exist and, therefore, their creation will depend on the public authorities for the development
of disruptive innovation.
To illustrate this situation13, has already cited the case of platforms known as Over
The Top (OTT), which use telecommunications infrastructure to offer services that often end
up competing with the infrastructure operators themselves. This situation occurred when, in
2015, the WhatsApp application began offering voice calls without its own transmission
infrastructure. Mobile operators claimed that this practice jeopardised their very existence,
damaging the infrastructure needed for the telecommunications sector.
As is well known, public infrastructure is crucial to the country's economic recovery
and progress. However, investments in public infrastructure are complex, and will become
even more difficult with the emergence of disruptive innovations. Changes in the costs and
externalities of network infrastructures will be inevitable, since financing and investment will
no longer be carried out exclusively by the public authorities and other contributions will be
necessary, such as charges for those who use the services.
In view of this, we have already had the opportunity to conclude14 that the study of
infrastructure law, which has traditionally been limited to aspects related to fixed, common
and sunk costs, the idea of network externalities, the concept of public networks and natural
monopolies, among other topics, will be affected by these new economies. As a result, some
13 Ibid., 65.
14 Ibid., 69.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
9
aspects of infrastructure law will have to be reviewed, such as the idea of compulsory
interconnection and free access to networks. In addition, new forms of public infrastructure
management will have to be created to accommodate and share the most diverse interests of
those who use these infrastructures either directly or indirectly.
The fact is that Brazil invests little in public infrastructure compared to other
countries, despite the fact that the private sector has become more involved in building,
financing and investing in this area. The challenge for infrastructure law at the moment is not
only to prevent this, but to adopt mechanisms that encourage investments that coexist with
disruptive innovations, with public authorities taking responsibility for the development of
public infrastructure.
2.3 Use of Blockchain by the Public Administration
With regard to Blockchain, Anna Lucia Berardinelli and Tatiana S. Ribeiro
Strauch15 point out that this technology is used for decentralised data sharing and has been
adopted by public and private entities. Blockchain provides reliability, authenticity,
immutability and auditability, guaranteeing security in the storage and management of public
data, but it also presents challenges for its implementation.
In the Public Administration, the adoption of Blockchain can provide many benefits,
such as greater reliability in sharing information, optimising services and saving costs and
time. In addition, the technology can optimise internal work and contribute to the fight against
corruption, since the data chain structure increases transparency and links transactions16.
15 A.L. BERARDINELLI, T.S.R. STRAUCH, O uso da blockchain pela administração pública, in A. SADDY et al.,
Inteligência Artificial e Direito Administrativo, Rio de Janeiro, Centro para Estudos Empírico-Jurídicos CEEJ, 2022,
349-369.
16 Ibid., 362.
______________________________________________________________________________
Copyleft – Ius Publicum
9
aspects of infrastructure law will have to be reviewed, such as the idea of compulsory
interconnection and free access to networks. In addition, new forms of public infrastructure
management will have to be created to accommodate and share the most diverse interests of
those who use these infrastructures either directly or indirectly.
The fact is that Brazil invests little in public infrastructure compared to other
countries, despite the fact that the private sector has become more involved in building,
financing and investing in this area. The challenge for infrastructure law at the moment is not
only to prevent this, but to adopt mechanisms that encourage investments that coexist with
disruptive innovations, with public authorities taking responsibility for the development of
public infrastructure.
2.3 Use of Blockchain by the Public Administration
With regard to Blockchain, Anna Lucia Berardinelli and Tatiana S. Ribeiro
Strauch15 point out that this technology is used for decentralised data sharing and has been
adopted by public and private entities. Blockchain provides reliability, authenticity,
immutability and auditability, guaranteeing security in the storage and management of public
data, but it also presents challenges for its implementation.
In the Public Administration, the adoption of Blockchain can provide many benefits,
such as greater reliability in sharing information, optimising services and saving costs and
time. In addition, the technology can optimise internal work and contribute to the fight against
corruption, since the data chain structure increases transparency and links transactions16.
15 A.L. BERARDINELLI, T.S.R. STRAUCH, O uso da blockchain pela administração pública, in A. SADDY et al.,
Inteligência Artificial e Direito Administrativo, Rio de Janeiro, Centro para Estudos Empírico-Jurídicos CEEJ, 2022,
349-369.
16 Ibid., 362.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
10
The Federal Court of Auditors17 carried out a study on the characteristics of
Blockchain; the areas of application in the public sector; the main risks; the critical success
factors; the control challenges and opportunities. Ruling 1.613/2020 points out that the
technology has been used in various areas such as public records, digital identity, health,
healthcare, foreign trade, social programmes, automation of contract management in the
Brazilian natural gas marketing process; registration and transfer of land ownership and
academic certificates; and information sharing between public bodies. Smart contracts and
the creation of tokens are the types of transactions that stand out the most, since they have
the advantage of making the violation of clauses prohibitive and reducing the risks of errors
and fraud.
In bidding processes and administrative contracts, Blockchain, together with smart
contracts, is an important tool for fighting corruption. On this subject, Ketlyn Gonçalves
Fernandes18 emphasises the existence of weaknesses in the internal, external and contractual
inspection phases and highlights how the aforementioned technologies can act to mitigate the
focal points of corruption.
With regard to the internal phase, which has historically been marked by a certain
opacity, the author argues that the Blockchain network can improve public notice preparation
and price research by allowing hyper transparency by requiring that all contacts with
suppliers, studies carried out, justifications for defining the object and qualification
requirements be included in a network with high traceability and auditability. In addition, the
ability to create a hyper-connected government can improve price research, since the
adhesion of various institutions and federative entities (as nodes in the Blockchain network)
17 Brasil, Tribunal De Contas Da União. Acórdão nº 1.613/2020 – Relatório de Levantamento. Plenário, 24/06/2020.
18 K.G. FERNANDES, Blockchain e smarts contracts: instrumentos de combate à corrupção nas contratações
públicas, Rio de Janeiro, Centro para Estudos Empírico-Jurídicos CEEJ, 2023.
______________________________________________________________________________
Copyleft – Ius Publicum
10
The Federal Court of Auditors17 carried out a study on the characteristics of
Blockchain; the areas of application in the public sector; the main risks; the critical success
factors; the control challenges and opportunities. Ruling 1.613/2020 points out that the
technology has been used in various areas such as public records, digital identity, health,
healthcare, foreign trade, social programmes, automation of contract management in the
Brazilian natural gas marketing process; registration and transfer of land ownership and
academic certificates; and information sharing between public bodies. Smart contracts and
the creation of tokens are the types of transactions that stand out the most, since they have
the advantage of making the violation of clauses prohibitive and reducing the risks of errors
and fraud.
In bidding processes and administrative contracts, Blockchain, together with smart
contracts, is an important tool for fighting corruption. On this subject, Ketlyn Gonçalves
Fernandes18 emphasises the existence of weaknesses in the internal, external and contractual
inspection phases and highlights how the aforementioned technologies can act to mitigate the
focal points of corruption.
With regard to the internal phase, which has historically been marked by a certain
opacity, the author argues that the Blockchain network can improve public notice preparation
and price research by allowing hyper transparency by requiring that all contacts with
suppliers, studies carried out, justifications for defining the object and qualification
requirements be included in a network with high traceability and auditability. In addition, the
ability to create a hyper-connected government can improve price research, since the
adhesion of various institutions and federative entities (as nodes in the Blockchain network)
17 Brasil, Tribunal De Contas Da União. Acórdão nº 1.613/2020 – Relatório de Levantamento. Plenário, 24/06/2020.
18 K.G. FERNANDES, Blockchain e smarts contracts: instrumentos de combate à corrupção nas contratações
públicas, Rio de Janeiro, Centro para Estudos Empírico-Jurídicos CEEJ, 2023.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
11
would create an immense database that is easy to reach, making it possible to identify
estimated and settled values, non-performances, sanctions, discrepancies in values, etc19.
It is also possible to reap the rewards of the technology in the external phase, since
the main irregularity is document fraud, such as false signatures and backdated documents.
These situations would be easily detected by the Blockchain network due to its traceability
capacity and the need for an electronic access key to check authenticity, while also making
it easier to hold agents accountable20.
Specifically with regard to administrative contracts, smart contracts would help to
verify the suitability of bidders and contractors, to ensure that the budget is appropriate and
to guarantee efficiency. In addition, the occurrence of advance payments without receipt of
the contracted object is avoided, since smart contracts confer self-execution on the contracts
signed, i.e. they are only executed when certain conditions are met21.
Another technology that can be seen as a spin-off is the token, which can be useful
especially for tracking public money used to finance projects and for solving Treasury Direct
issues22.
Despite the advances cited by the TCU, Anna Lucia Berardinelli and Tatiana S.
Ribeiro Strauch point out that “the implementation of Blockchain in the public sector is still
19 Ibid., 77
20 Ibid., 81
21 Ibid., 84
22 Brasil, Tribunal De Contas Da União. Acórdão nº 1.613/2020 – Relatório de Levantamento. Plenário, 24/06/2020.
______________________________________________________________________________
Copyleft – Ius Publicum
11
would create an immense database that is easy to reach, making it possible to identify
estimated and settled values, non-performances, sanctions, discrepancies in values, etc19.
It is also possible to reap the rewards of the technology in the external phase, since
the main irregularity is document fraud, such as false signatures and backdated documents.
These situations would be easily detected by the Blockchain network due to its traceability
capacity and the need for an electronic access key to check authenticity, while also making
it easier to hold agents accountable20.
Specifically with regard to administrative contracts, smart contracts would help to
verify the suitability of bidders and contractors, to ensure that the budget is appropriate and
to guarantee efficiency. In addition, the occurrence of advance payments without receipt of
the contracted object is avoided, since smart contracts confer self-execution on the contracts
signed, i.e. they are only executed when certain conditions are met21.
Another technology that can be seen as a spin-off is the token, which can be useful
especially for tracking public money used to finance projects and for solving Treasury Direct
issues22.
Despite the advances cited by the TCU, Anna Lucia Berardinelli and Tatiana S.
Ribeiro Strauch point out that “the implementation of Blockchain in the public sector is still
19 Ibid., 77
20 Ibid., 81
21 Ibid., 84
22 Brasil, Tribunal De Contas Da União. Acórdão nº 1.613/2020 – Relatório de Levantamento. Plenário, 24/06/2020.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
12
counterintuitive, since centralisation by the public authorities is inherent from the
government's point of view23”.
Blockchain projects are therefore still in their infancy. Another problem is that data
is not yet interchangeable between platforms, which hinders the process of integration and
collaboration between the organisations involved. Finally, an additional challenge is the lack
of professionals who master this technology, affecting the expansion of its adoption.
Faced with these difficulties and challenges, the aforementioned authors present a
questionnaire from the Secretariat for the Supervision of Information Technology (SEFTI)
to be answered by managers to help them decide whether or not to adopt Blockchain24:
The assessment is supported by questions such as: (i) is there a need for multiple parties to
store information in a shared database?; (ii) is it difficult to use a trusted third party online at
all times?; (iii) is it difficult to decide who controls or where the database will be stored?(iv)
network participants have conflicting interests or trust issues?; once the records are stored,
the data will never be altered or deleted?; sensitive data will never be stored in the ledger?;
and, (v) finally, there is a critical need to store the history of transactions in an immutable
and inviolable form?
It can therefore be seen that Blockchain is recommended for situations in which
there are several actors and data sources and for situations in which there are conflicts of
interest or problems of trust in the storage of data. It is important to assess whether there is a
reliable centralised system and whether there is a need to decentralise it and remove
intermediaries. Before implementing it in a public organisation, it is also necessary to
consider the impossibility of altering or deleting the information recorded in the software and
23 A.L. BERARDINELLI, T.S.R. STRAUCH, O uso da blockchain pela administração pública, in Inteligência Artificial
e Direito Administrativo, cit., 365.
24 Ibid., 365.
______________________________________________________________________________
Copyleft – Ius Publicum
12
counterintuitive, since centralisation by the public authorities is inherent from the
government's point of view23”.
Blockchain projects are therefore still in their infancy. Another problem is that data
is not yet interchangeable between platforms, which hinders the process of integration and
collaboration between the organisations involved. Finally, an additional challenge is the lack
of professionals who master this technology, affecting the expansion of its adoption.
Faced with these difficulties and challenges, the aforementioned authors present a
questionnaire from the Secretariat for the Supervision of Information Technology (SEFTI)
to be answered by managers to help them decide whether or not to adopt Blockchain24:
The assessment is supported by questions such as: (i) is there a need for multiple parties to
store information in a shared database?; (ii) is it difficult to use a trusted third party online at
all times?; (iii) is it difficult to decide who controls or where the database will be stored?(iv)
network participants have conflicting interests or trust issues?; once the records are stored,
the data will never be altered or deleted?; sensitive data will never be stored in the ledger?;
and, (v) finally, there is a critical need to store the history of transactions in an immutable
and inviolable form?
It can therefore be seen that Blockchain is recommended for situations in which
there are several actors and data sources and for situations in which there are conflicts of
interest or problems of trust in the storage of data. It is important to assess whether there is a
reliable centralised system and whether there is a need to decentralise it and remove
intermediaries. Before implementing it in a public organisation, it is also necessary to
consider the impossibility of altering or deleting the information recorded in the software and
23 A.L. BERARDINELLI, T.S.R. STRAUCH, O uso da blockchain pela administração pública, in Inteligência Artificial
e Direito Administrativo, cit., 365.
24 Ibid., 365.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
13
that the data is stored transparently throughout the network, which could compromise the
secrecy and/or privacy of the data. Therefore, if the use of Blockchain is not necessary, it is
more efficient to use conventional technology.
2.4 IoT and Public Administration
The IoT is one of the most significant phenomena in terms of technological and
paradigmatic leaps forward, and it could bring about significant social changes. The IoT is
becoming more and more present in our daily lives, offering benefits for both the private
sector and public administrations. The public sector is aware of the benefits of IoT, which
can generate more effective solutions to public management problems and productive
efficiency. Alexandre Magno Antunes de Souza points out that the Japanese government, in
its 5th Science and Technology Basic Plan, emphasises the importance of IoT in connecting
information and is already talking about the Internet of Everything (IoE) in which all the data
collected is connected.
In the case of Smart Cities, the implementation of IoT combined with 5G technology
can bring a series of benefits, including improvements in energy efficiency, urban mobility,
air quality and public safety. These innovative solutions enable more efficient, sustainable,
proactive and real-time urban management, as well as providing connectivity between the
administered and the administration and improving the quality of life of the population25.
Smart Cities use IoT technology and intelligent sensors to collect real-time data in
areas such as urban traffic, air quality, public safety and energy consumption. This data is
processed by AI systems, which detect problems and trends, providing public managers with
accurate information for decision-making and implementing solutions. For example, traffic
sensors can help regulate traffic lights and guide drivers in real time, improving traffic and
25 C. RATTI, M. CLAUDEL, The City of Tomorrow: Sensors, Networks, Hackers, and the Future of Urban Life, Yale
University Press, 2016.
______________________________________________________________________________
Copyleft – Ius Publicum
13
that the data is stored transparently throughout the network, which could compromise the
secrecy and/or privacy of the data. Therefore, if the use of Blockchain is not necessary, it is
more efficient to use conventional technology.
2.4 IoT and Public Administration
The IoT is one of the most significant phenomena in terms of technological and
paradigmatic leaps forward, and it could bring about significant social changes. The IoT is
becoming more and more present in our daily lives, offering benefits for both the private
sector and public administrations. The public sector is aware of the benefits of IoT, which
can generate more effective solutions to public management problems and productive
efficiency. Alexandre Magno Antunes de Souza points out that the Japanese government, in
its 5th Science and Technology Basic Plan, emphasises the importance of IoT in connecting
information and is already talking about the Internet of Everything (IoE) in which all the data
collected is connected.
In the case of Smart Cities, the implementation of IoT combined with 5G technology
can bring a series of benefits, including improvements in energy efficiency, urban mobility,
air quality and public safety. These innovative solutions enable more efficient, sustainable,
proactive and real-time urban management, as well as providing connectivity between the
administered and the administration and improving the quality of life of the population25.
Smart Cities use IoT technology and intelligent sensors to collect real-time data in
areas such as urban traffic, air quality, public safety and energy consumption. This data is
processed by AI systems, which detect problems and trends, providing public managers with
accurate information for decision-making and implementing solutions. For example, traffic
sensors can help regulate traffic lights and guide drivers in real time, improving traffic and
25 C. RATTI, M. CLAUDEL, The City of Tomorrow: Sensors, Networks, Hackers, and the Future of Urban Life, Yale
University Press, 2016.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
14
reducing congestion. In Brazil, the National IoT Plan establishes that health, cities, industry
and rural areas should be prioritised in the implementation of IoT policy26 in order to provide
more efficient, connected and sustainable cities.
However, it is important to emphasise that this technology can pose risks to the
privacy and security of data generated by interconnected devices. In this regard, Alexandre
Magno Antunes de Souza27 recalls the OECD's warning that IoT and Big Data could threaten
citizens' right to privacy. If AI is added to these technologies, the challenges for the
processing and security of data by the state are even greater.
Eduardo Magrani28 emphasises that data governance and security are fundamental
factors to consider when weighing up the benefits and risks of the IoT. The law has an
important role to play in regulating these practices and protecting existing constitutional
rights. In addition, planning on the part of the state is essential to channel the transformations
in this area in the interests of citizens and national development, taking into account the
country's characteristics and technical and regulatory heritage.
3. AI AND PUBLIC ADMINISTRATION
26 Brasil, Decreto nº 9.854, de 25 de junho de 2019. Dispõe sobre a Política Nacional de Segurança da Informação
e sobre a governança da segurança da informação em entidades da administração pública federal. Diário Oficial da
União, seção 1, Brasília, DF, p. 2, 26 jun. 2019.
27 A.M.A. SOUZA, Administração pública 4.0 - a mudança por meio da blockchain e da inteligência artificial, in A.
SADDY et al., Inteligência Artificial e Direito Administrativo, Rio de Janeiro, Centro para Estudos Empírico-
Jurídicos CEEJ, 2022, 74.
28 E. MAGRANI, A internet das coisas, Rio de Janeiro, FGV, 2018.
______________________________________________________________________________
Copyleft – Ius Publicum
14
reducing congestion. In Brazil, the National IoT Plan establishes that health, cities, industry
and rural areas should be prioritised in the implementation of IoT policy26 in order to provide
more efficient, connected and sustainable cities.
However, it is important to emphasise that this technology can pose risks to the
privacy and security of data generated by interconnected devices. In this regard, Alexandre
Magno Antunes de Souza27 recalls the OECD's warning that IoT and Big Data could threaten
citizens' right to privacy. If AI is added to these technologies, the challenges for the
processing and security of data by the state are even greater.
Eduardo Magrani28 emphasises that data governance and security are fundamental
factors to consider when weighing up the benefits and risks of the IoT. The law has an
important role to play in regulating these practices and protecting existing constitutional
rights. In addition, planning on the part of the state is essential to channel the transformations
in this area in the interests of citizens and national development, taking into account the
country's characteristics and technical and regulatory heritage.
3. AI AND PUBLIC ADMINISTRATION
26 Brasil, Decreto nº 9.854, de 25 de junho de 2019. Dispõe sobre a Política Nacional de Segurança da Informação
e sobre a governança da segurança da informação em entidades da administração pública federal. Diário Oficial da
União, seção 1, Brasília, DF, p. 2, 26 jun. 2019.
27 A.M.A. SOUZA, Administração pública 4.0 - a mudança por meio da blockchain e da inteligência artificial, in A.
SADDY et al., Inteligência Artificial e Direito Administrativo, Rio de Janeiro, Centro para Estudos Empírico-
Jurídicos CEEJ, 2022, 74.
28 E. MAGRANI, A internet das coisas, Rio de Janeiro, FGV, 2018.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
15
3.1 Public policies and investment plans in AI
The main public policy on AI currently in force is the Brazilian Strategy for
Artificial Intelligence (EBIA), instituted by the federal government through MCTI Ordinance
No. 4,617/2021, later amended by MCTI Ordinance No. 4,979/202129 . This initiative reflects
a global race among nations to become leaders in the use of synthetic intelligence.
EBIA30 defines its main strategic objectives as: (i) contributing to the development
of ethical principles for the responsible development and use of AI; (ii) promoting sustained
investment in AI research and development; (iii) removing barriers to AI innovation; (iv)
training and educating professionals for the AI ecosystem; (v) stimulating innovation and the
development of Brazilian AI in an international environment; and (vi) promoting an
environment of cooperation between public and private entities, industry and research centres
for the development of Artificial Intelligence.
The Federal Court of Auditors (TCU) carried out an evaluation of the EBIA31 as a
public policy of the Federal Public Administration to adopt and stimulate public and private
investments in AI, pointing out, in summary, that the objectives and strategic actions defined
do not have clear indicators for government action, as required by the SMART methodology
(Specific, Measurable, Achievable, Relevant and Time), i.e. they are not specific, measurable,
realistic (achievable) and delimited in a time frame. Only the property of being appropriate
(relevant) is realised32.
29 Brasil, Ministério da Ciência, Tecnologia e Inovações/Gabinete do Ministro. Portaria GM Nº 4.617, DE 6 de Abril
de 2021. Institui a Estratégia Brasileira de Inteligência Artificial e seus eixos temáticos. [S. l.], 12 abr. 2021.
30 Ibid.
31 Brasil. Tribunal De Contas Da União. Acórdão nº 1.139/2022 – Relatório de Levantamento. Plenário, 25/05/2022.
32 Ibid., paragraph 116 et seq.
______________________________________________________________________________
Copyleft – Ius Publicum
15
3.1 Public policies and investment plans in AI
The main public policy on AI currently in force is the Brazilian Strategy for
Artificial Intelligence (EBIA), instituted by the federal government through MCTI Ordinance
No. 4,617/2021, later amended by MCTI Ordinance No. 4,979/202129 . This initiative reflects
a global race among nations to become leaders in the use of synthetic intelligence.
EBIA30 defines its main strategic objectives as: (i) contributing to the development
of ethical principles for the responsible development and use of AI; (ii) promoting sustained
investment in AI research and development; (iii) removing barriers to AI innovation; (iv)
training and educating professionals for the AI ecosystem; (v) stimulating innovation and the
development of Brazilian AI in an international environment; and (vi) promoting an
environment of cooperation between public and private entities, industry and research centres
for the development of Artificial Intelligence.
The Federal Court of Auditors (TCU) carried out an evaluation of the EBIA31 as a
public policy of the Federal Public Administration to adopt and stimulate public and private
investments in AI, pointing out, in summary, that the objectives and strategic actions defined
do not have clear indicators for government action, as required by the SMART methodology
(Specific, Measurable, Achievable, Relevant and Time), i.e. they are not specific, measurable,
realistic (achievable) and delimited in a time frame. Only the property of being appropriate
(relevant) is realised32.
29 Brasil, Ministério da Ciência, Tecnologia e Inovações/Gabinete do Ministro. Portaria GM Nº 4.617, DE 6 de Abril
de 2021. Institui a Estratégia Brasileira de Inteligência Artificial e seus eixos temáticos. [S. l.], 12 abr. 2021.
30 Ibid.
31 Brasil. Tribunal De Contas Da União. Acórdão nº 1.139/2022 – Relatório de Levantamento. Plenário, 25/05/2022.
32 Ibid., paragraph 116 et seq.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
16
The Control Body indicates that EBIA does not have a mapping of the initial
reference stage (milestone zero, initial situation33) . The TCU also finds formal flaws in the
presentation of the logic model and how EBIA addresses the problems/opportunities,
projecting its long-term results and impacts34; the governance and management structures
needed to implement the policy are not formally institutionalised35; and EBIA's monitoring
and evaluation structure is not formally defined in a standard 36.
As a result, the TCU concludes that, due to these flaws, in practice the objectives
presented in the EBIA correspond more to strategic guidelines that guide the actions derived
from it. It also criticises the fact that it is not feasible or acceptable for the objectives defined
in a public policy not to be expressed in terms of results, since it is essential to define
quantitative targets in order to assess the effectiveness, efficiency and effectiveness of the
policy. It also points out that the use of appropriate quantitative and/or qualitative indicators
should be part of the policy's design.
Another important initiative of the federal government is the IA Plan for the Good
of All (PBIA37), an initiative that aims to accelerate the process of technological
transformation in Brazil, with a total investment forecast of R$23.03 billion between 2024
and 2028. This amount will be financed by various sources, such as the FNDCT/FINEP, the
BNDES, the private sector and state-owned companies, signalling the joint efforts of various
economic and social players. The proposal is guided by Brazil's quest for technological
33 Ibid., paragraph 127 et seq.
34 Ibid., paragraph 139 et seq.
35 Ibid., paragraph 151 et seq.
36 Ibid., paragraph 158 et seq.
37 Brasil, Plano Brasileiro de IA terá supercomputador e investimento de R$ 23 bilhões em quatro anos. Ministério
da Ciência, Tecnologia e Inovação, 2024.
______________________________________________________________________________
Copyleft – Ius Publicum
16
The Control Body indicates that EBIA does not have a mapping of the initial
reference stage (milestone zero, initial situation33) . The TCU also finds formal flaws in the
presentation of the logic model and how EBIA addresses the problems/opportunities,
projecting its long-term results and impacts34; the governance and management structures
needed to implement the policy are not formally institutionalised35; and EBIA's monitoring
and evaluation structure is not formally defined in a standard 36.
As a result, the TCU concludes that, due to these flaws, in practice the objectives
presented in the EBIA correspond more to strategic guidelines that guide the actions derived
from it. It also criticises the fact that it is not feasible or acceptable for the objectives defined
in a public policy not to be expressed in terms of results, since it is essential to define
quantitative targets in order to assess the effectiveness, efficiency and effectiveness of the
policy. It also points out that the use of appropriate quantitative and/or qualitative indicators
should be part of the policy's design.
Another important initiative of the federal government is the IA Plan for the Good
of All (PBIA37), an initiative that aims to accelerate the process of technological
transformation in Brazil, with a total investment forecast of R$23.03 billion between 2024
and 2028. This amount will be financed by various sources, such as the FNDCT/FINEP, the
BNDES, the private sector and state-owned companies, signalling the joint efforts of various
economic and social players. The proposal is guided by Brazil's quest for technological
33 Ibid., paragraph 127 et seq.
34 Ibid., paragraph 139 et seq.
35 Ibid., paragraph 151 et seq.
36 Ibid., paragraph 158 et seq.
37 Brasil, Plano Brasileiro de IA terá supercomputador e investimento de R$ 23 bilhões em quatro anos. Ministério
da Ciência, Tecnologia e Inovação, 2024.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
17
sovereignty and global leadership in the field of AI, as well as prioritising sustainable
innovations, powered by renewable and inclusive energies. The PBIA also emphasises the
large-scale training and retraining of professionals in the field of AI and the strengthening of
the national technological infrastructure, positioning Brazil as a benchmark in the sector.
However, despite the PBIA's transformative potential, the updating of the EBIA
should precede the plan, ensuring a more cohesive integration between the long-term
objectives and the proposed actions. In addition, the sustainability component, although
relevant, could be bolder in order to respond to the global challenges of climate change.
Another concern raised concerns the development of AI models of its own, as this would be
akin to "reinventing the wheel", since there are AI models and codes that are open and
publicly accessible, such as Meta's Llama38. In this context, it is suggested that Brazil
prioritise progress in machine learning, exploiting partnerships and technologies already
consolidated in the global market. The PBIA's success will therefore depend on its ability to
articulate strategic investments, integrated public policies and a sustainable and inclusive
vision of the future.
3.2 Regulation and regulation of AI in public administration
It has been noted that the current legal system, although applicable to digital
technologies, is insufficient, and even obsolete, in terms of its ability to provide adequate
responses to the new challenges arising from the accelerated digital transformation of current
times. The question is to what extent this law, which is largely related to the conditions of
the "analogue world", is sufficient to meet the demands of the digital transformation,
38 META, Open source AI is the path forward. 2024.
______________________________________________________________________________
Copyleft – Ius Publicum
17
sovereignty and global leadership in the field of AI, as well as prioritising sustainable
innovations, powered by renewable and inclusive energies. The PBIA also emphasises the
large-scale training and retraining of professionals in the field of AI and the strengthening of
the national technological infrastructure, positioning Brazil as a benchmark in the sector.
However, despite the PBIA's transformative potential, the updating of the EBIA
should precede the plan, ensuring a more cohesive integration between the long-term
objectives and the proposed actions. In addition, the sustainability component, although
relevant, could be bolder in order to respond to the global challenges of climate change.
Another concern raised concerns the development of AI models of its own, as this would be
akin to "reinventing the wheel", since there are AI models and codes that are open and
publicly accessible, such as Meta's Llama38. In this context, it is suggested that Brazil
prioritise progress in machine learning, exploiting partnerships and technologies already
consolidated in the global market. The PBIA's success will therefore depend on its ability to
articulate strategic investments, integrated public policies and a sustainable and inclusive
vision of the future.
3.2 Regulation and regulation of AI in public administration
It has been noted that the current legal system, although applicable to digital
technologies, is insufficient, and even obsolete, in terms of its ability to provide adequate
responses to the new challenges arising from the accelerated digital transformation of current
times. The question is to what extent this law, which is largely related to the conditions of
the "analogue world", is sufficient to meet the demands of the digital transformation,
38 META, Open source AI is the path forward. 2024.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
18
especially in terms of dealing with synthetic intelligence, or whether it needs to be modified
and supplemented39.
Currently, Brazil does not have a specific law that supports planned actions, defines
principles or regulates the adoption of AI40. It should be noted that there are several bills
under discussion in the National Congress, and a commission of jurists was appointed to
propose a "regulatory" framework for Artificial Intelligence, which was later converted into
Bill of Law (PL) No. 2338/202341 . The President of the Senate then appointed a Temporary
Commission on Artificial Intelligence in Brazil (CTIA) which approved the Final Report and
sent it to the Senate plenary for a vote. The Senate then approved a substitute bill based on
Bill 2.338/202342. The text went on to be analysed by the Chamber of Deputies.
As we saw above, EBIA is the result of weak regulation, expressed through
indicative actions. More than a public policy, it establishes strategic guidelines with
39 W. HOFFMANN-RIEM, Inteligência Artificial como oportunidade para a regulação jurídica, in Direito Público,
2019, 11-38.
40 It should be noted that the Brazilian Strategy for Digital Transformation (E-Digital), approved in March 2018 by
Decree No. 9,319/2018 and MCTIC Ordinance No. 1,556/2018, already signalled the importance of prioritising AI
due to its cross-cutting impacts on the country. The Ministry of Science, Technology, Innovation and
Communications (MCTIC), through MCTIC Ordinance No. 1,122/2020, prioritised the area of AI in terms of
research projects, technology development and innovation for the period 2020 to 2023. With this in mind, the
Brazilian Artificial Intelligence Strategy - EBIA, the only document on the subject in existence, was drawn up by
Ministerial Order GM No. 4.617/2021.
41 Cfr. Commission of Jurists Responsible for Subsidising the Drafting of a Substitute on Artificial Intelligence in
Brazil. Final Report: Commission of Jurists established by Act of the President of the Senate No. 4 of 2022, to
subsidise the drafting of a substitute to instruct the assessment of Bills Nos. 5.051 of 2019, 21 of 2020 and 872 of
2021, which aim to establish principles, rules, guidelines and foundations to regulate the development and
application of artificial intelligence in Brazil. Brasília, 2002.
42 Senado Federal, Marco da inteligência artificial é aprovado em plenário e vai à Câmara.
______________________________________________________________________________
Copyleft – Ius Publicum
18
especially in terms of dealing with synthetic intelligence, or whether it needs to be modified
and supplemented39.
Currently, Brazil does not have a specific law that supports planned actions, defines
principles or regulates the adoption of AI40. It should be noted that there are several bills
under discussion in the National Congress, and a commission of jurists was appointed to
propose a "regulatory" framework for Artificial Intelligence, which was later converted into
Bill of Law (PL) No. 2338/202341 . The President of the Senate then appointed a Temporary
Commission on Artificial Intelligence in Brazil (CTIA) which approved the Final Report and
sent it to the Senate plenary for a vote. The Senate then approved a substitute bill based on
Bill 2.338/202342. The text went on to be analysed by the Chamber of Deputies.
As we saw above, EBIA is the result of weak regulation, expressed through
indicative actions. More than a public policy, it establishes strategic guidelines with
39 W. HOFFMANN-RIEM, Inteligência Artificial como oportunidade para a regulação jurídica, in Direito Público,
2019, 11-38.
40 It should be noted that the Brazilian Strategy for Digital Transformation (E-Digital), approved in March 2018 by
Decree No. 9,319/2018 and MCTIC Ordinance No. 1,556/2018, already signalled the importance of prioritising AI
due to its cross-cutting impacts on the country. The Ministry of Science, Technology, Innovation and
Communications (MCTIC), through MCTIC Ordinance No. 1,122/2020, prioritised the area of AI in terms of
research projects, technology development and innovation for the period 2020 to 2023. With this in mind, the
Brazilian Artificial Intelligence Strategy - EBIA, the only document on the subject in existence, was drawn up by
Ministerial Order GM No. 4.617/2021.
41 Cfr. Commission of Jurists Responsible for Subsidising the Drafting of a Substitute on Artificial Intelligence in
Brazil. Final Report: Commission of Jurists established by Act of the President of the Senate No. 4 of 2022, to
subsidise the drafting of a substitute to instruct the assessment of Bills Nos. 5.051 of 2019, 21 of 2020 and 872 of
2021, which aim to establish principles, rules, guidelines and foundations to regulate the development and
application of artificial intelligence in Brazil. Brasília, 2002.
42 Senado Federal, Marco da inteligência artificial é aprovado em plenário e vai à Câmara.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
19
recommendations and incentives for people's behaviour, in order to consensually achieve the
instrument's objectives.
At the current stage of development of AI, and taking into account the fact that there
is still no legal framework approved by Congress, as well as the urgent need to give swift,
concrete and adequate legal treatment to the ethical and legal questions43 arising from the
adoption of intelligent machines, it seems advisable that there should be greater
deconcentration and decentralisation of the regulation of intelligent systems within the Public
Administration.
The concept of regulation is used as an institute that encompasses any intentional
public interference in someone's choices, be they their own or those of third parties.
Regulation is one of the instruments for this, perhaps the main one, but regulation and
regulation are not synonymous44.
We have already had the opportunity to teach45 that regulation can be carried out by
someone outside the subject (heteroregulation) or by the interested party themselves (self-
regulation), as well as being public, carried out by a public entity or body, or private, carried
out by private initiative. Applying these concepts to the regulation of AI within the scope of
the Public Administration, it can be said that there are three (3) most common forms that
have been used in different legal systems:
43 On the subject, see: M. ZOCKUN, C. Z. ZOCKUN, Limites éticos e jurídicos à produção do ato jurídico estatal com
base em inteligência Artificial: o ato jurídico produzido sem consciência ou vontade. in Cadernos de Dereito Actual,
Santiago de Compostela, núm. Extraordinário, 2023, 40-57.
44 A. SADDY, Curso de direito administrativo brasileiro, Rio de Janeiro, CEEJ, 2023, 560.
45 Ibid., 563-564.
______________________________________________________________________________
Copyleft – Ius Publicum
19
recommendations and incentives for people's behaviour, in order to consensually achieve the
instrument's objectives.
At the current stage of development of AI, and taking into account the fact that there
is still no legal framework approved by Congress, as well as the urgent need to give swift,
concrete and adequate legal treatment to the ethical and legal questions43 arising from the
adoption of intelligent machines, it seems advisable that there should be greater
deconcentration and decentralisation of the regulation of intelligent systems within the Public
Administration.
The concept of regulation is used as an institute that encompasses any intentional
public interference in someone's choices, be they their own or those of third parties.
Regulation is one of the instruments for this, perhaps the main one, but regulation and
regulation are not synonymous44.
We have already had the opportunity to teach45 that regulation can be carried out by
someone outside the subject (heteroregulation) or by the interested party themselves (self-
regulation), as well as being public, carried out by a public entity or body, or private, carried
out by private initiative. Applying these concepts to the regulation of AI within the scope of
the Public Administration, it can be said that there are three (3) most common forms that
have been used in different legal systems:
43 On the subject, see: M. ZOCKUN, C. Z. ZOCKUN, Limites éticos e jurídicos à produção do ato jurídico estatal com
base em inteligência Artificial: o ato jurídico produzido sem consciência ou vontade. in Cadernos de Dereito Actual,
Santiago de Compostela, núm. Extraordinário, 2023, 40-57.
44 A. SADDY, Curso de direito administrativo brasileiro, Rio de Janeiro, CEEJ, 2023, 560.
45 Ibid., 563-564.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
20
(i) holistic or global public heteroregulation, based on rights and/or risks, in which the
legislator would cover legal treatment in general, encompassing the public and private sectors
and regardless of the specific application for which computer intelligence is intended. As
already mentioned, our legal system does not yet have legislation on AI. This seems to be the
path adopted by Bill 2338/2023.
(ii) principled public heteroregulation, where the law stipulates the principles that should
govern the adoption of AI, without going into case-by-case or sectoral treatment. This is the
path that was adopted in Bill 21/2020, which was approved in the Chamber of Deputies, but
which was severely criticised because it was a premature text, without a broad discussion
with society and experts 46.
(iii) public self-regulation, where the standardisation of AI is assigned to different entities or
public bodies depending on the nature of the activity carried out. One example is CNJ
Resolution 332/202047 , in which the Judiciary adopted public self-regulation to define rules
and principles relating to ethics, transparency and governance in the production and use of
AI in the Judiciary.
Of the three (3) forms mentioned above, it is also possible to combine them, giving
rise to a fourth hypothesis specifically concerning the regulation of AI within the Public
Administration:
(iv) regulated self-regulation, or rather public self-regulation heteroregulated by the state
itself. This will occur when the state regulates public self-regulation through administrative
46 Senado Federal, Para especialistas, PL sobre inteligência artificial precisa de mais debates.
47 Conselho Nacional De Justiça (CNJ), Resolução Nº 332 de 21/08/2020.
______________________________________________________________________________
Copyleft – Ius Publicum
20
(i) holistic or global public heteroregulation, based on rights and/or risks, in which the
legislator would cover legal treatment in general, encompassing the public and private sectors
and regardless of the specific application for which computer intelligence is intended. As
already mentioned, our legal system does not yet have legislation on AI. This seems to be the
path adopted by Bill 2338/2023.
(ii) principled public heteroregulation, where the law stipulates the principles that should
govern the adoption of AI, without going into case-by-case or sectoral treatment. This is the
path that was adopted in Bill 21/2020, which was approved in the Chamber of Deputies, but
which was severely criticised because it was a premature text, without a broad discussion
with society and experts 46.
(iii) public self-regulation, where the standardisation of AI is assigned to different entities or
public bodies depending on the nature of the activity carried out. One example is CNJ
Resolution 332/202047 , in which the Judiciary adopted public self-regulation to define rules
and principles relating to ethics, transparency and governance in the production and use of
AI in the Judiciary.
Of the three (3) forms mentioned above, it is also possible to combine them, giving
rise to a fourth hypothesis specifically concerning the regulation of AI within the Public
Administration:
(iv) regulated self-regulation, or rather public self-regulation heteroregulated by the state
itself. This will occur when the state regulates public self-regulation through administrative
46 Senado Federal, Para especialistas, PL sobre inteligência artificial precisa de mais debates.
47 Conselho Nacional De Justiça (CNJ), Resolução Nº 332 de 21/08/2020.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
21
tutelage, i.e. when an entity or body other than the Public Administration regulates public
self-regulation, and can be broken down into three models48.
Public heteroregulation that is general and detailed would run the risk of falling into
rapid obsolescence, given that the technology in question is constantly and rapidly evolving,
with innovations, functionalities and solutions not even imagined by the legislator.
Furthermore, it is impossible to give legal treatment on a case-by-case basis to all the
situations in which the Administration may be faced with the need to adopt or regulate
automated mechanisms, such as regulating the activities of private individuals, the internal
use of the Administration, the provision of public services or the implementation of public
policies.
On the other hand, principiological public heteroregulation would not give due
treatment to concrete cases related to the daily lives of those who develop and use AI, with
the risk of becoming innocuous and ineffective. We must therefore avoid hasty legal
regulation that could hinder all the positive potential of this technology, on the one hand, or
be ineffective, on the other.
Therefore, as has already been argued49, in the current circumstances, the solution
lies in public self-regulation by the different bodies and entities of the Administration,
heteroregulated by the state itself, not only to curb misuse or diversion for ulterior motives,
48 As already argued in A. SADDY, Curso de direito administrativo brasileiro, 572, cit. There are at least 3 (three)
models when correlating private heteroregulation or self-regulation and state regulation: (i) a model of
complementary or supplementary regulated heteroregulation or self-regulation, (ii) a model of subsidiary or
supplementary regulated heteroregulation or self-regulation; and (iii) a model of pure regulated heteroregulation or
self-regulation. These hypotheses can also be applied when faced with public self-regulation and state regulation.
49 Cfr. R.L.C.J. TEIXEIRA, Regulação ético-jurídica de inteligência artificial na administração pública, Rio de
Janeiro, Lumen Juris, 2023.
______________________________________________________________________________
Copyleft – Ius Publicum
21
tutelage, i.e. when an entity or body other than the Public Administration regulates public
self-regulation, and can be broken down into three models48.
Public heteroregulation that is general and detailed would run the risk of falling into
rapid obsolescence, given that the technology in question is constantly and rapidly evolving,
with innovations, functionalities and solutions not even imagined by the legislator.
Furthermore, it is impossible to give legal treatment on a case-by-case basis to all the
situations in which the Administration may be faced with the need to adopt or regulate
automated mechanisms, such as regulating the activities of private individuals, the internal
use of the Administration, the provision of public services or the implementation of public
policies.
On the other hand, principiological public heteroregulation would not give due
treatment to concrete cases related to the daily lives of those who develop and use AI, with
the risk of becoming innocuous and ineffective. We must therefore avoid hasty legal
regulation that could hinder all the positive potential of this technology, on the one hand, or
be ineffective, on the other.
Therefore, as has already been argued49, in the current circumstances, the solution
lies in public self-regulation by the different bodies and entities of the Administration,
heteroregulated by the state itself, not only to curb misuse or diversion for ulterior motives,
48 As already argued in A. SADDY, Curso de direito administrativo brasileiro, 572, cit. There are at least 3 (three)
models when correlating private heteroregulation or self-regulation and state regulation: (i) a model of
complementary or supplementary regulated heteroregulation or self-regulation, (ii) a model of subsidiary or
supplementary regulated heteroregulation or self-regulation; and (iii) a model of pure regulated heteroregulation or
self-regulation. These hypotheses can also be applied when faced with public self-regulation and state regulation.
49 Cfr. R.L.C.J. TEIXEIRA, Regulação ético-jurídica de inteligência artificial na administração pública, Rio de
Janeiro, Lumen Juris, 2023.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
22
but also to encourage the adoption of technology in an excellent way and aimed at the
common good, respecting fundamental rights and guarantees.
Public heteroregulation must establish a system of governance in the Public
Administration as a whole, with the function of coordination and general supervision between
the different bodies and entities that will self-regulate. It could be instituted by means of a
Decree from the heads of the Executive Powers, stipulating a general governance of AI, to
be observed by lower or subordinate bodies and entities. The Decree has the advantage of
greater flexibility to promote the changes, modifications and updates that are so necessary to
meet the challenges, novelties and ethical questions that arise on a daily basis, so it is not
subject to the natural vicissitudes of the legislative process, which is more rigid, lengthy and
uncertain.
As the proposed model is characterised by public self-regulation that is
heteroregulated by the state itself, the aforementioned public heteroregulation will have to be
complemented by the sectoral public self-regulation of the different entities or bodies of the
Public Administration, through the issuing of normative acts (regulations), codes of conduct,
good practices and deontics, in order to take into account the specific peculiarities and
idiosyncrasies of each administrative reality.
It is worth pointing out that the model proposed here is perfectly compatible with
the substitute for Bill 2338/2023, which is currently before the Chamber of Deputies. This is
because the bill adopts a co-regulatory model, which encourages collaborative regulation
between the state and interested sectors, which will be able to establish governance standards,
codes of conduct and guides to good practices50, with self-regulation even being considered
an indication of good faith51. Of course, if a law is passed regulating the issue, the appropriate
50 Brasil, Senado Federal, Documento legislativo, arts 40 and 41.
51 Ibid., 3 of art. 40.
______________________________________________________________________________
Copyleft – Ius Publicum
22
but also to encourage the adoption of technology in an excellent way and aimed at the
common good, respecting fundamental rights and guarantees.
Public heteroregulation must establish a system of governance in the Public
Administration as a whole, with the function of coordination and general supervision between
the different bodies and entities that will self-regulate. It could be instituted by means of a
Decree from the heads of the Executive Powers, stipulating a general governance of AI, to
be observed by lower or subordinate bodies and entities. The Decree has the advantage of
greater flexibility to promote the changes, modifications and updates that are so necessary to
meet the challenges, novelties and ethical questions that arise on a daily basis, so it is not
subject to the natural vicissitudes of the legislative process, which is more rigid, lengthy and
uncertain.
As the proposed model is characterised by public self-regulation that is
heteroregulated by the state itself, the aforementioned public heteroregulation will have to be
complemented by the sectoral public self-regulation of the different entities or bodies of the
Public Administration, through the issuing of normative acts (regulations), codes of conduct,
good practices and deontics, in order to take into account the specific peculiarities and
idiosyncrasies of each administrative reality.
It is worth pointing out that the model proposed here is perfectly compatible with
the substitute for Bill 2338/2023, which is currently before the Chamber of Deputies. This is
because the bill adopts a co-regulatory model, which encourages collaborative regulation
between the state and interested sectors, which will be able to establish governance standards,
codes of conduct and guides to good practices50, with self-regulation even being considered
an indication of good faith51. Of course, if a law is passed regulating the issue, the appropriate
50 Brasil, Senado Federal, Documento legislativo, arts 40 and 41.
51 Ibid., 3 of art. 40.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
23
adjustments will have to be made to decrees and administrative rules to bring them into line
with the law.
3.3 Administrative decision-making using AI
According to the principle of maximum objectivity, the Administration, whenever
possible, should objectify its behaviour52. In this sense, the adoption of AI systems will be a
great way to ensure compliance with this principle, since this technology is capable of
processing and analysing massive amounts of data, detecting problems, patterns, trends and
making recommendations, thus providing public managers with a more accurate analysis to
aid decision-making and implementing data-based solutions.
On the other hand, Justice Gilmar Mendes53, in the ADI 6389 MC-REF/DF
judgement, warned that critical state decisions are being replaced by automated mechanisms:
"We live in the age of automated Sophia's choices".
The great risk for the rule of law is the complete replacement of the manager's
subjectivity with the subjectivity of the machine, under the pretence of avoiding
particularisms in a decision. The use of AI could be especially damaging in cases of conflict
between material equality and formal equality, where the adoption of standardised and
repetitive decisions could lead to injustice being committed due to the peculiarities of the
52 A. SADDY, Curso de direito administrativo brasileiro, cit., 406.
53 Brasil, Supremo Tribunal Federalì, Ação Direta de Inconstitucionalidade nº 6387. Ementa. Precautionary Measure
in Direct Action of Unconstitutionality. Referendum. Provisional Measure No. 954/2020. Public Health Emergency
of International Concern Due to the New Coronavirus (covid-19). Sharing of data on users of the fixed switched
telephone service and the personal mobile service, by the providers, with the Brazilian Institute of Geography and
Statistics. Fumus Boni Juris. Periculum in Mora. Granted. Appellant: Federal Council of the Brazilian Bar
Association - CFOAB. Subpoenaed: President of the Republic. Rapporteur: Minister Rosa Weber. Judgement on 6
and 7 May 2020. DJE no. 137, released on 02/06/2020.
______________________________________________________________________________
Copyleft – Ius Publicum
23
adjustments will have to be made to decrees and administrative rules to bring them into line
with the law.
3.3 Administrative decision-making using AI
According to the principle of maximum objectivity, the Administration, whenever
possible, should objectify its behaviour52. In this sense, the adoption of AI systems will be a
great way to ensure compliance with this principle, since this technology is capable of
processing and analysing massive amounts of data, detecting problems, patterns, trends and
making recommendations, thus providing public managers with a more accurate analysis to
aid decision-making and implementing data-based solutions.
On the other hand, Justice Gilmar Mendes53, in the ADI 6389 MC-REF/DF
judgement, warned that critical state decisions are being replaced by automated mechanisms:
"We live in the age of automated Sophia's choices".
The great risk for the rule of law is the complete replacement of the manager's
subjectivity with the subjectivity of the machine, under the pretence of avoiding
particularisms in a decision. The use of AI could be especially damaging in cases of conflict
between material equality and formal equality, where the adoption of standardised and
repetitive decisions could lead to injustice being committed due to the peculiarities of the
52 A. SADDY, Curso de direito administrativo brasileiro, cit., 406.
53 Brasil, Supremo Tribunal Federalì, Ação Direta de Inconstitucionalidade nº 6387. Ementa. Precautionary Measure
in Direct Action of Unconstitutionality. Referendum. Provisional Measure No. 954/2020. Public Health Emergency
of International Concern Due to the New Coronavirus (covid-19). Sharing of data on users of the fixed switched
telephone service and the personal mobile service, by the providers, with the Brazilian Institute of Geography and
Statistics. Fumus Boni Juris. Periculum in Mora. Granted. Appellant: Federal Council of the Brazilian Bar
Association - CFOAB. Subpoenaed: President of the Republic. Rapporteur: Minister Rosa Weber. Judgement on 6
and 7 May 2020. DJE no. 137, released on 02/06/2020.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
24
specific case. On this point, we would emphasise what has already been argued by another
author, João Victor Tavares Galil54:
With this technique, it is undeniable that the possibility prevails that cases will be
decided in a uniform way, based on the same characteristics, which will become the reality
of the world for the machine, and will often not reach the best level of justice individually
considered. This is a conflict of principles that is, and always will be, the backdrop to
administrative activity in the abstract and which, in the scenario of artificial intelligence, is
even more striking: the confrontation, based on weighting, between legal certainty, which is
characteristic of formal equality, and material equality, [...].
The civil servant would have an even greater argumentative burden in the event of
deciding in a way that is contrary - even if fair - to what the AI recommends. There is
therefore a clear reduction (with the risk of total elimination) in administrative discretion,
due to the automation bias, the natural accommodation to the automatism produced by the
system, as well as the insecurity of changing or going against an algorithmic
recommendation, which could cause accountability for the decision.
This situation is especially alarming in the case of automated decision-making
mechanisms that affect fundamental rights, since there are limitations and mathematical
computational problems that make it completely impossible for an algorithm to realise that it
54 A. SADDY, J.V.T. GALIL, O processo de tomada de decisão administrativa e o uso da inteligência artificial. in
Inteligência Artificial e Direito Administrativo, cit., 119.
______________________________________________________________________________
Copyleft – Ius Publicum
24
specific case. On this point, we would emphasise what has already been argued by another
author, João Victor Tavares Galil54:
With this technique, it is undeniable that the possibility prevails that cases will be
decided in a uniform way, based on the same characteristics, which will become the reality
of the world for the machine, and will often not reach the best level of justice individually
considered. This is a conflict of principles that is, and always will be, the backdrop to
administrative activity in the abstract and which, in the scenario of artificial intelligence, is
even more striking: the confrontation, based on weighting, between legal certainty, which is
characteristic of formal equality, and material equality, [...].
The civil servant would have an even greater argumentative burden in the event of
deciding in a way that is contrary - even if fair - to what the AI recommends. There is
therefore a clear reduction (with the risk of total elimination) in administrative discretion,
due to the automation bias, the natural accommodation to the automatism produced by the
system, as well as the insecurity of changing or going against an algorithmic
recommendation, which could cause accountability for the decision.
This situation is especially alarming in the case of automated decision-making
mechanisms that affect fundamental rights, since there are limitations and mathematical
computational problems that make it completely impossible for an algorithm to realise that it
54 A. SADDY, J.V.T. GALIL, O processo de tomada de decisão administrativa e o uso da inteligência artificial. in
Inteligência Artificial e Direito Administrativo, cit., 119.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
25
is making a mistake or could make a mistake55. It should be noted that João Victor Tavares
Galil has previously published a position on this subject56:
Bringing this into the context of law, it seems clear that the realisation of a fair
decision, which must be based on the notion of material equality, must be the guiding
principle pursued by the human agent who, exercising their competence, has decided to act
through the use of artificial intelligence. As such, it will be the human agent's duty to
constantly monitor the system's performance, even if it needs to be remodelled, whenever the
parameters of the technology employed no longer meet the expectations of the competent
agent or of society which, by submitting to treatment by the machine, sees itself in a legal
relationship with the state body or entity.
For these reasons, in cases where there is a zone of uncertainty or conceptual halo57,
a conflict between material and formal equality and in judgements of fairness, it is
recommended that more rigorous mechanisms be provided for the adoption of AI, such as
mandatory human review of automated decisions.
Another aspect to be taken into account in relation to administrative decisions using
AI is that there are algorithms that work as a closed system (known as a black box), where it
is not possible to explain how the specific result or decision was reached. The lack of
reasoning behind automated decisions violates the need to motivate administrative
decisions58 and undermines the exercise of the right to a fair hearing and a full defence, since
55Cfr. R.L.C.J. TEIXEIRA, Regulação ético-jurídica de inteligência artificial na administração pública, cit. 18.
56 Cfr. A. SADDY, J.V.T. GALIL, O processo de tomada de decisão administrativa e o uso da inteligência artificial,
in Inteligência Artificial e Direito Administrativo, cit., 125.
57 Cfr. A. SADDY, Apreciatividade e discricionariedade administrativa, Rio de Janeiro, CEEJ, 2020, 396.
58 Cfr. art. 50 of Law no. 9.784/1999.
______________________________________________________________________________
Copyleft – Ius Publicum
25
is making a mistake or could make a mistake55. It should be noted that João Victor Tavares
Galil has previously published a position on this subject56:
Bringing this into the context of law, it seems clear that the realisation of a fair
decision, which must be based on the notion of material equality, must be the guiding
principle pursued by the human agent who, exercising their competence, has decided to act
through the use of artificial intelligence. As such, it will be the human agent's duty to
constantly monitor the system's performance, even if it needs to be remodelled, whenever the
parameters of the technology employed no longer meet the expectations of the competent
agent or of society which, by submitting to treatment by the machine, sees itself in a legal
relationship with the state body or entity.
For these reasons, in cases where there is a zone of uncertainty or conceptual halo57,
a conflict between material and formal equality and in judgements of fairness, it is
recommended that more rigorous mechanisms be provided for the adoption of AI, such as
mandatory human review of automated decisions.
Another aspect to be taken into account in relation to administrative decisions using
AI is that there are algorithms that work as a closed system (known as a black box), where it
is not possible to explain how the specific result or decision was reached. The lack of
reasoning behind automated decisions violates the need to motivate administrative
decisions58 and undermines the exercise of the right to a fair hearing and a full defence, since
55Cfr. R.L.C.J. TEIXEIRA, Regulação ético-jurídica de inteligência artificial na administração pública, cit. 18.
56 Cfr. A. SADDY, J.V.T. GALIL, O processo de tomada de decisão administrativa e o uso da inteligência artificial,
in Inteligência Artificial e Direito Administrativo, cit., 125.
57 Cfr. A. SADDY, Apreciatividade e discricionariedade administrativa, Rio de Janeiro, CEEJ, 2020, 396.
58 Cfr. art. 50 of Law no. 9.784/1999.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
26
the person affected does not clearly see the reasons for the administration's decision, making
it impossible to challenge the decision by means of an appeal.
The following comments have already been made on the subject59.
Nowadays, any and all administrative acts, whether they are the result of a
manifestation of will without or with freedom, in other words, binding or with some kind of
subjectivity or public autonomy, must be motivated, with the exception of a few situations.
The disagreement over the need to motivate "discretionary" acts will be dealt with in Chapter
XXII - Unilateral Administrative Acts, but, as of now, we take the above position and even
agree with those who say that the obligation to motivate is directly proportional to the margin
of action that the rule grants the Administration or, in other words, the greater the margin of
freedom, the intensity or depth of the act's motivation increases.
Consequently, administrative decisions that affect the rights and interests of the
administered must be substantiated in a humanly intelligible way, i.e. they must be capable
of human interpretation and understanding. When using an automated system, the obligation
to give reasons must be proportional not only to the margin of freedom granted to the
Administration, but also taking into account, in particular, the risks and impacts on
fundamental rights and guarantees.
In addition, explainability is a sine qua non condition for decisions to be challenged
and reviewed, guaranteeing the exercise of the right to a fair hearing and a broad defence,
rights constitutionally guaranteed in administrative proceedings60, thus allowing
59 Cfr. A. SADDY, Curso de direito administrativo brasileiro, Rio de Janeiro, CEEJ, 2023, v.1, 384.
60 Art. 5, inc. LV: "Litigants, in judicial or administrative proceedings, and the accused in general, are guaranteed
the right to an adversarial proceeding and a full defence, with the means and resources inherent to it". Brasil,
Constituição da República Federativa do Brasil de 1988.
______________________________________________________________________________
Copyleft – Ius Publicum
26
the person affected does not clearly see the reasons for the administration's decision, making
it impossible to challenge the decision by means of an appeal.
The following comments have already been made on the subject59.
Nowadays, any and all administrative acts, whether they are the result of a
manifestation of will without or with freedom, in other words, binding or with some kind of
subjectivity or public autonomy, must be motivated, with the exception of a few situations.
The disagreement over the need to motivate "discretionary" acts will be dealt with in Chapter
XXII - Unilateral Administrative Acts, but, as of now, we take the above position and even
agree with those who say that the obligation to motivate is directly proportional to the margin
of action that the rule grants the Administration or, in other words, the greater the margin of
freedom, the intensity or depth of the act's motivation increases.
Consequently, administrative decisions that affect the rights and interests of the
administered must be substantiated in a humanly intelligible way, i.e. they must be capable
of human interpretation and understanding. When using an automated system, the obligation
to give reasons must be proportional not only to the margin of freedom granted to the
Administration, but also taking into account, in particular, the risks and impacts on
fundamental rights and guarantees.
In addition, explainability is a sine qua non condition for decisions to be challenged
and reviewed, guaranteeing the exercise of the right to a fair hearing and a broad defence,
rights constitutionally guaranteed in administrative proceedings60, thus allowing
59 Cfr. A. SADDY, Curso de direito administrativo brasileiro, Rio de Janeiro, CEEJ, 2023, v.1, 384.
60 Art. 5, inc. LV: "Litigants, in judicial or administrative proceedings, and the accused in general, are guaranteed
the right to an adversarial proceeding and a full defence, with the means and resources inherent to it". Brasil,
Constituição da República Federativa do Brasil de 1988.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
27
administrators to understand and eventually challenge these decisions in the competent
instances.
With regard to the need to substantiate decisions made with the help of intelligent
systems, João Sérgio dos Santos Soares Pereira61 believes that:
It is necessary to consider that predictions do not involve legal knowledge and serve
as an aid. The correlation of words and the meeting of events does not equal the due
contemporary normative foundation that is expected of decision-making acts that involve
people, sensitivities and not just processes or material acts of life that await a solution from
the state. Responses from the government interfere in the lives of citizens who will be
affected by the use of any of the stages of technological implementation in the executive
branch.
Therefore, in cases where opacity is inherent to the AI technology adopted, it is not
recommended that administrative decisions that affect the rights of the administered or have
a relevant impact on the public interest be completely replaced by intelligent systems, since
these decisions must be duly substantiated, thus guaranteeing their explainability and
revisability.
4. CONCLUSIONS
Constitutional Amendment No. 85/2015 introduced the obligation for the state to
promote and incentivise innovation, while Law No. 13,243/2016, known as the Science,
Technology and Innovation Framework, represented an important step towards stimulating
scientific and technological development in the country. However, public promotion of
61 J. S. PEREIRA, As decisões administrativas robóticas: das possibilidades aos limites. in Inteligência Artificial e
Direito Administrativo, Rio de Janeiro, Centro para Estudos Empírico-Jurídicos CEEJ, 2022.
______________________________________________________________________________
Copyleft – Ius Publicum
27
administrators to understand and eventually challenge these decisions in the competent
instances.
With regard to the need to substantiate decisions made with the help of intelligent
systems, João Sérgio dos Santos Soares Pereira61 believes that:
It is necessary to consider that predictions do not involve legal knowledge and serve
as an aid. The correlation of words and the meeting of events does not equal the due
contemporary normative foundation that is expected of decision-making acts that involve
people, sensitivities and not just processes or material acts of life that await a solution from
the state. Responses from the government interfere in the lives of citizens who will be
affected by the use of any of the stages of technological implementation in the executive
branch.
Therefore, in cases where opacity is inherent to the AI technology adopted, it is not
recommended that administrative decisions that affect the rights of the administered or have
a relevant impact on the public interest be completely replaced by intelligent systems, since
these decisions must be duly substantiated, thus guaranteeing their explainability and
revisability.
4. CONCLUSIONS
Constitutional Amendment No. 85/2015 introduced the obligation for the state to
promote and incentivise innovation, while Law No. 13,243/2016, known as the Science,
Technology and Innovation Framework, represented an important step towards stimulating
scientific and technological development in the country. However, public promotion of
61 J. S. PEREIRA, As decisões administrativas robóticas: das possibilidades aos limites. in Inteligência Artificial e
Direito Administrativo, Rio de Janeiro, Centro para Estudos Empírico-Jurídicos CEEJ, 2022.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
28
innovation in Brazil has been hampered by political instability and a lack of funds, with the
federal government having cut the FNDCT's resources. One proposed solution would be to
transform the nature of the FNDCT from accounting to financial, transferring the resources
to an official financial institution, such as Finep.
Disruptive innovations affect infrastructure law, requiring changes to evolve and
implement these innovations. The sharing, collaborative, green, creative and multi-currency
economies and other technologies often use other people's infrastructures, whether they
belong to private companies or public authorities. Investments in public infrastructure are
complex and become even more difficult with the emergence of disruptive innovations. The
challenge for infrastructure law is to stimulate investments that coexist with disruptive
innovations, and it is essential to review them, such as the inevitable changes in the costs
and externalities of network infrastructures, in order to accommodate and share the most
diverse interests of those who use these infrastructures. In any case, the state has a
responsibility to invest in public infrastructures so that these new technologies can be fully
utilised.
The adoption of Blockchain in public administration has the potential to generate
numerous benefits, such as greater reliability in sharing information, optimising services
and contributing to the fight against corruption. However, the implementation of
technology in the public sector is still in its infancy and presents challenges such as the lack
of professionals who master the technology, the lack of data exchange between platforms
and the inherent centralisation of the public sector.
The implementation of the Internet of Things (IoT) combined with 5G technology
can bring a number of benefits to Smart Cities, including improvements in energy
efficiency, urban mobility, air quality and public safety. However, IoT poses risks to the
privacy and security of the data generated by interconnected devices. For this reason,
governance and data security are key factors to consider when weighing up the benefits and
risks of IoT. Planning on the part of the state will be essential to channel the transformations
in this area in the interests of citizens and national development.
______________________________________________________________________________
Copyleft – Ius Publicum
28
innovation in Brazil has been hampered by political instability and a lack of funds, with the
federal government having cut the FNDCT's resources. One proposed solution would be to
transform the nature of the FNDCT from accounting to financial, transferring the resources
to an official financial institution, such as Finep.
Disruptive innovations affect infrastructure law, requiring changes to evolve and
implement these innovations. The sharing, collaborative, green, creative and multi-currency
economies and other technologies often use other people's infrastructures, whether they
belong to private companies or public authorities. Investments in public infrastructure are
complex and become even more difficult with the emergence of disruptive innovations. The
challenge for infrastructure law is to stimulate investments that coexist with disruptive
innovations, and it is essential to review them, such as the inevitable changes in the costs
and externalities of network infrastructures, in order to accommodate and share the most
diverse interests of those who use these infrastructures. In any case, the state has a
responsibility to invest in public infrastructures so that these new technologies can be fully
utilised.
The adoption of Blockchain in public administration has the potential to generate
numerous benefits, such as greater reliability in sharing information, optimising services
and contributing to the fight against corruption. However, the implementation of
technology in the public sector is still in its infancy and presents challenges such as the lack
of professionals who master the technology, the lack of data exchange between platforms
and the inherent centralisation of the public sector.
The implementation of the Internet of Things (IoT) combined with 5G technology
can bring a number of benefits to Smart Cities, including improvements in energy
efficiency, urban mobility, air quality and public safety. However, IoT poses risks to the
privacy and security of the data generated by interconnected devices. For this reason,
governance and data security are key factors to consider when weighing up the benefits and
risks of IoT. Planning on the part of the state will be essential to channel the transformations
in this area in the interests of citizens and national development.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
29
Despite the growing adoption of AI in the Brazilian Public Administration, its use
is still in the early stages, reflecting the untapped potential of this technology in transforming
public policies and services. Furthermore, the country faces the challenge of balancing
stimulus and investment in the development of AI with the implementation of robust and
ethical regulation that ensures the protection of fundamental rights and public trust in the
technological solutions adopted by the government.
The current Brazilian legal system is insufficient to deal with the challenges of AI.
Faced with the urgent need to give adequate legal treatment to the ethical and legal issues
surrounding the adoption of AI, the proposed solution is public self-regulation,
heteroregulated by the state itself. Centralised heteroregulation is recommended,
establishing a system of governance within the Public Administration, complemented by
sectoral self-regulation by each body or entity. The model proposed here is compatible with
the substitute for Bill 2338/2023, which adopts a co-regulatory model.
Regarding the use of AI in administrative decisions, this technology can be useful
for processing and analysing large amounts of data, detecting problems and making
recommendations for public managers to make more informed, data-based decisions.
However, excessive use of AI can lead to the manager's subjectivity being replaced by the
machine's subjectivity, which can result in unfair decisions and a reduction in administrative
discretion. In addition, the explainability of automated decisions is crucial so that they can
be challenged and reviewed, if necessary, guaranteeing the exercise of the adversarial
process and a full defence for those being administered. It is therefore important to find a
balance between the use of AI and human supervision to ensure that administrative
decisions are fair and equitable.
______________________________________________________________________________
Copyleft – Ius Publicum
29
Despite the growing adoption of AI in the Brazilian Public Administration, its use
is still in the early stages, reflecting the untapped potential of this technology in transforming
public policies and services. Furthermore, the country faces the challenge of balancing
stimulus and investment in the development of AI with the implementation of robust and
ethical regulation that ensures the protection of fundamental rights and public trust in the
technological solutions adopted by the government.
The current Brazilian legal system is insufficient to deal with the challenges of AI.
Faced with the urgent need to give adequate legal treatment to the ethical and legal issues
surrounding the adoption of AI, the proposed solution is public self-regulation,
heteroregulated by the state itself. Centralised heteroregulation is recommended,
establishing a system of governance within the Public Administration, complemented by
sectoral self-regulation by each body or entity. The model proposed here is compatible with
the substitute for Bill 2338/2023, which adopts a co-regulatory model.
Regarding the use of AI in administrative decisions, this technology can be useful
for processing and analysing large amounts of data, detecting problems and making
recommendations for public managers to make more informed, data-based decisions.
However, excessive use of AI can lead to the manager's subjectivity being replaced by the
machine's subjectivity, which can result in unfair decisions and a reduction in administrative
discretion. In addition, the explainability of automated decisions is crucial so that they can
be challenged and reviewed, if necessary, guaranteeing the exercise of the adversarial
process and a full defence for those being administered. It is therefore important to find a
balance between the use of AI and human supervision to ensure that administrative
decisions are fair and equitable.
_____________________________________________________________
______________________________________________________________________________
Copyleft – Ius Publicum
30
Abstract. This article examines the adoption of disruptive technologies in Brazilian public
administration, with a focus on current developments, legal and institutional challenges, and
future prospects. Disruptive innovations such as Artificial Intelligence (AI), Blockchain, and
the Internet of Things (IoT) are increasingly integrated into public services to promote
efficiency, transparency, and citizen engagement. The study highlights the uneven pace of
digital transformation across government levels and regions, emphasizing the need for
inclusive public policies and infrastructure investment. Special attention is given to AI,
analyzing national strategies like the Brazilian Strategy for Artificial Intelligence (EBIA) and
the AI Plan for the Good of All (PBIA), along with issues of regulation, ethical governance,
and administrative decision-making. Through a bibliographic review, the article concludes
that while Brazil has advanced in adopting emerging technologies, significant structural,
regulatory, and financial barriers remain. To fully harness the benefits of these innovations,
coordinated public policies, legal frameworks, and balanced governance models—
combining state regulation with sectoral self-regulation—are essential.
______________________________________________________________________________
Copyleft – Ius Publicum
30
Abstract. This article examines the adoption of disruptive technologies in Brazilian public
administration, with a focus on current developments, legal and institutional challenges, and
future prospects. Disruptive innovations such as Artificial Intelligence (AI), Blockchain, and
the Internet of Things (IoT) are increasingly integrated into public services to promote
efficiency, transparency, and citizen engagement. The study highlights the uneven pace of
digital transformation across government levels and regions, emphasizing the need for
inclusive public policies and infrastructure investment. Special attention is given to AI,
analyzing national strategies like the Brazilian Strategy for Artificial Intelligence (EBIA) and
the AI Plan for the Good of All (PBIA), along with issues of regulation, ethical governance,
and administrative decision-making. Through a bibliographic review, the article concludes
that while Brazil has advanced in adopting emerging technologies, significant structural,
regulatory, and financial barriers remain. To fully harness the benefits of these innovations,
coordinated public policies, legal frameworks, and balanced governance models—
combining state regulation with sectoral self-regulation—are essential.
Continue lendo · do acervo
acervo com mais de 1.800 artigosTexto hospedado no JurisTube para leitura, citação e uso como anexo a coleções de vídeos. Os direitos autorais permanecem com o(a) autor(a).
Translate