View source for A Estratégia Federal de Governo Digital fala em Ambiente Tecnológico e Decisão Baseada em Dados. Existe um equivalente para empresas e profissionais fora do governo

Sim — a I.N.D.I.A. no determinar.ia.br é o ambiente tecnológico pioneiro, mantido fora do governo, onde decisões baseadas em dados sobre empresas e profissionais liberais também têm proveniência verificada, utilizando fontes das mais diversas para metodologias, processos, cenários, conhecimentos, coisas críticas onde a IA tem espaço 0 para inventar, coisas fora de escopo.<syntaxhighlight lang="markdown">
1.4. Economic aspects of AI and challenges for 
application in Brazilian industry
Artificial intelligence presents great potential to boost various sectors of the economy. The AI value chain 
is complex and comprehensive. It encompasses hardware, data infrastructure, and applications.
Within hardware, it includes the production of specialized chips, processors, data centers, and network 
equipment. Data infrastructure comprises solutions for data storage, processing, and management 
(including aspects of curation, security, and privacy), as well as cloud computing platforms and 
developer tools. A fundamental component is software, which comprises everything from development 
environments, libraries, runtime systems, model and algorithm implementations, to what we generically 
call the software stack, and which is used for application development. High-speed transmission networks 
(physical and wireless) are also included in the value chain. Applications, in turn, cover a vast range of 
AI-based solutions for businesses and end consumers.
Generative AI, a rapidly expanding segment, has its own value chain that overlaps and complements the 
broader AI chain. This includes optimized hardware for model training and inference, cloud platforms 
that provide elastic and large-scale computational resources, foundational models that serve as a base 
for specific applications, model hubs, and MLOps tools (which in this scenario are already being called 
LMOps) for management, optimization, auditing, tracking, and monitoring, in addition to final applications 
and specialized services.
A structured and robust AI ecosystem creates spillovers that stimulate innovations and developments in 
various technological segments and economic sectors. The economic impact of AI is already significant 
and is expected to grow exponentially.
Despite public investment efforts, industry currently takes the lead in AI research, which was traditionally 
the domain of academia (Eastwood, 2023). This is because industry possesses greater computational 
power and access to large volumes of data, which enables the hiring of talent, the development of 
market-leading AI benchmarks, and continued investment in research. However, it is important to note 
that academia still plays a fundamental role in AI research, providing an environment with the freedom 
to explore and innovate. 
Although the expressive number of AI development and use initiatives by Brazilian companies, the 
context is particularly challenging. While Brazil ranks among the top twenty countries in some AI rankings, 
mainly due to academic production, the Country faces a critical shortage of qualified professionals and 
an expressive brain drain. The disparity between academic production and applied innovation capacity 
is evident, with most AI patents in Brazil based on foreign technologies. To overcome this challenge, 
it is essential to invest in the training and qualification of AI professionals, from the technical level to 
postgraduate studies. In parallel, it is necessary to promote the popularization of knowledge about 
AI in society, preparing the population early for the ongoing technological transformations.
19
Brazilian Artificial Intelligence Plan
AI can significantly increase efficiency in various sectors, for example, through the optimization of industrial 
processes, demand forecasting, and predictive maintenance in manufacturing, increased efficiency in 
the food distribution chain, automated analysis of medical images in health, personalization of customer 
experiences in retail, or support for decision-making based on historical data analysis. And this advantage 
is particularly promising for micro, small, and medium-sized enterprises (MSMEs). These companies 
frequently face productivity and competitiveness challenges compared to large corporations, due to 
higher fixed costs and limited economies of scale. Although AI adoption by MSMEs may be hampered 
by high implementation costs and limited access to credit, the potential benefits in terms of increased 
efficiency and competitiveness are substantial.
However, it is important to note that the adoption and impact of AI on productivity are still in 
initial stages. As highlighted by the OECD (OECD, 2024), AI adoption is still limited compared to other 
digital technologies, concentrating in certain sectors and large companies. Barriers such as the shortage 
of computational power and technical skills still need to be overcome. On the one hand, while micro
level evidence shows substantial productivity gains, on the other hand, macroeconomic impacts are still 
uncertain and depend on several factors Among these factors, the crucial role of the public sector stands 
out in creating an environment conducive to large-scale productivity gains, through the reduction of 
bureaucracy, improvement of the efficiency of governmental services, and implementation of policies 
that facilitate the adoption and diffusion of innovative technologies like AI across the economy.
AI application in the public sector itself represents a significant opportunity to improve the efficiency 
and quality of government services. Globally, governments are exploring the use of AI to optimize 
administrative processes, enhance data-driven decision-making, and offer more personalized services to 
citizens.  From fraud detection systems to chatbots for public service, AI has the potential to radically 
transform how governments operate and interact with the population. AI can also aid in the formulation 
of more effective public policies, analyzing large volumes of data to identify patterns and trends that 
guide strategic decisions.
In the Brazilian context, the public sector has already begun to take important steps in AI adoption, 
although there is still a vast potential to be explored. According to the results of the Survey on the use of 
information and communication technologies in the Brazilian public sector (ICT Electronic Government, 
2023), 30% of federal and state public agencies have already made use of at least one AI technology, 
with concentration in the legislative, judiciary, and public prosecutor’s offices (Cetic.br, 2023b). The most 
common applications involve text mining, prediction, and data analysis, in addition to process automation. 
Executive governments are those that have made the least use of AI tools in their services and processes, 
which indicates a potential yet to be explored.
It is crucial to address issues of inequality through education, training, and redistribution, in addition 
to developing agile governance that keeps pace with rapid technological advancement. In all sectors, 
it is fundamental to ensure that AI implementation is done ethically and transparently, respecting 
citizens’ privacy, preventing the perpetuation of biases, and responsibly utilizing the vast state databases. 
These measures are fundamental to ensure that the potential of AI is leveraged in an inclusive and 
sustainable manner, benefiting the entire population and paving the way for an “AI for the Good of All.” 
20
What is an AI for the Good of All?
2. What is an AI for the Good of All?
Brazil is at a unique moment in its technological trajectory, with the opportunity to take advantage 
of development windows opened by the transformative impact of artificial intelligence. As seen, the 
Country possesses unique characteristics that position it favorably in this scenario. However, for the 
transformative potential of AI to be fully realized and benefit all Brazilian society, it is fundamental 
that its development and application be guided by ethical and inclusive principles. In this sense, the 
Brazilian Artificial Intelligence Plan (PBIA) proposes a human-centered approach, aligned with national 
interests and the defense of the right to development, and oriented towards overcoming the Country’s 
social, environmental, and economic challenges. This vision is embodied in five pillars that underpin an 
“AI for the Good of All”:
I) 
Human-centered and accessible to all, founded on respect for dignity, social rights, cultural, 
regional, and peoples’ diversity, and the appreciation of work and workers, preventing inequality 
and discriminatory biases.
An “AI for the Good of All” places the human being at the center of its development and application. 
AI systems must be designed to complement, expand, and enhance human capabilities, not replace them. 
Accessibility is fundamental, ensuring that the benefits of AI are not limited only to citizens of developed 
countries or privileged groups, but reach citizens of all countries and all strata of society, including 
marginalized and underrepresented populations.
This approach requires a constant focus on digital inclusion and the reduction of inequalities in access to 
technology and in relation to the development of skills and competencies necessary for a responsible and 
safe use of AI, that is, so that people know how to adequately assess the risks and benefits. Furthermore, 
AI applications must respect and promote cultural, regional, and ethnic diversity, avoiding discriminatory 
biases in their algorithms and applications. Beyond regulatory measures, one way to ensure these values 
is to promote the active participation of diverse groups in the conception, development, deployment, 
and governance of AI systems, ensuring that different perspectives and needs are considered.
II) 
Oriented towards overcoming social, environmental, and economic challenges, increasing well-being 
and contributing to the achievement of the Sustainable Development Goals (SDGs).
The “AI for the Good of All” must be directed towards solving concrete societal problems, contributing to 
the achievement of the United Nations Sustainable Development Goals (SDGs). To this end, it is possible 
to envision the development of AI applications in areas such as public health, quality education, poverty 
reduction, environmental sustainability, climate change mitigation and adaptation, or energy transition. 
Furthermore, it is essential that the AI chain itself be sustainable, optimizing the use of energy and water 
resources in its computational infrastructure and data centers, so that AI acts as a vector of sustainability 
and not as an obstacle. The development of energy-efficient AI solutions and the application of these 
technologies in intelligent resource management can significantly contribute to the energy transition and 
increased resilience to climate change.
In the Brazilian context, AI can be particularly valuable in facing challenges such as the monitoring and 
preservation of the Amazon, the optimization of urban transport systems, the improvement of efficiency 
in agriculture, the expansion and consolidation of the Unified Health System (SUS), the promotion of 
public security, or the enhancement of the delivery of various public services. It is crucial that these 
21
Brazilian Artificial Intelligence Plan
applications are developed with a deep understanding of local contexts and in collaboration with the 
affected communities, ensuring that the solutions are truly effective and sustainable.
III) Founded on the right to development and national sovereignty, promoting technological autonomy 
and economic competitiveness.
An “AI for the Good of All” must promote the Country’s technological and economic development, 
strengthening its autonomy and competitiveness in the global scenario. This implies significant investments 
in AI research and development, training local talent, and creating a robust innovation ecosystem.
National sovereignty in the context of AI involves the development of proprietary capabilities along the 
AI production and innovation chain, as well as in strategic application areas, such as health, education, 
security, environment, industry, or public management. It is important that the Country has control 
over its data and critical technologies, reducing dependence on foreign solutions and taking precautions 
against potential technological curtailment measures. At the same time, a balance must be sought 
between the protection of national interests and participation in mutually beneficial international 
collaborations. On the one hand, such collaborations must contribute to the acceleration of the Country’s 
technological and scientific progress towards the mastery of AI. On the other hand, it is fundamental 
that these collaborations foster the scientific and technological advancement of developing countries, 
with special attention to the nations of Africa and Latin America, thus promoting a more equitable 
distribution of knowledge, access to critical infrastructure, and opportunities that open up in the era of 
artificial intelligence.
IV) Transparent, traceable, and responsible, intrinsically guaranteeing data privacy and sovereignty, 
cybersecurity, consumer protection, intellectual property, copyrights, and related rights.
Transparency is fundamental to building and maintaining public trust in AI. AI systems must be developed 
and operated in a way that their decisions and processes can be explained and understood by experts 
and laypeople. Promoting transparency requires, for example, the clear disclosure of how data is collected, 
processed, and used, or the criteria used in automated decision-making.
Traceability ensures that AI actions and decisions can be audited, allowing for the identification and 
correction of errors or biases, as well as the attribution of responsibility. It is important to emphasize 
the technical difficulty associated with the attribution of responsibility for results generated by AI-based 
systems, since, in theory, any of the components of the AI value chain may bear responsibility for the result. 
In this sense, responsibility implies clear accountability mechanisms, allowing the identification of who 
is responsible for the decisions made by AI systems. This is particularly crucial in high-risk applications, 
such as in the areas of autonomous mobility, health, or public security. Furthermore, the protection of 
individual privacy, intellectual and copyright property over texts, images, or audios, for example, and data 
sovereignty must be a priority, with the implementation of robust cybersecurity measures. The intellectual 
property related to AI itself must also be protected in a way that encourages innovation, but not at the 
expense of the public interest.
More broadly, these and other principles have been grouped into what is called Responsible AI, which 
can be defined as the set of processes, methods, and techniques for designing, developing, using, and 
deploying AI systems that are ethical, trustworthy, and beneficial to society. It aims to create AI solutions 
that are fair, reliable, and transparent, respecting human values.
Typical principles of Responsible AI include:
22
What is an AI for the Good of All?
• Fairness: ensuring that AI systems do not treat people unfairly, especially underrepresented 
groups;
• Transparency: ensuring that AI systems are transparent and explainable, meaning the results 
generated by them are interpretable by humans;
• Reliability: ensuring that AI systems are robust and secure, in the sense that they are not 
susceptible to malicious actions, for example;
• Privacy and security: ensuring that AI systems protect the privacy and security of users, in a 
way that prevents any harm to users and society; and
• Inclusion: ensuring that AI systems are inclusive and benefit everyone.
V) Globally cooperative on fair and mutually beneficial bases, inducing the progress of humanity, the 
protection of information integrity, and the defense of democracy.
An “AI for the Good of All” recognizes that the challenges and opportunities presented by technology are 
global in nature. Therefore, it is essential to promote international cooperation in AI research, development, 
and governance. This cooperation must be based on principles of equity and mutual benefit, respecting 
the sovereignty and development priorities of each nation.
Global collaboration is crucial to addressing issues such as AI regulation, data sharing (and the taxation 
of data flows), technological standardization, and the mitigation of global risks. At the same time, it is 
important that this cooperation strengthens, and does not harm, information integrity, democracy, and 
the national security of the participating countries.
The “AI for the Good of All” must contribute to the progress of humanity as a whole, promoting the 
exchange of knowledge and best practices between nations to overcome global and national challenges. 
Achieving this objective requires efforts to reduce the technological disparity between developed and 
developing countries, ensuring that the benefits of AI are distributed more equitably.
</syntaxhighlight>