Brazilian legislative proposals for Artificial Intelligence

In PL 21/2020, other relevant elements should be stressed, such as, for example, the scope of the definitions of “AI agents” and “interested parties”.

In the Chamber of Deputies (Draft Laws No. 21/2020 and 240/2020) and in the Federal Senate (Draft Laws No. 5051/2019 and 5691/2019), proposals for rules aimed at, in summary, stipulating principles, guidelines, guidelines and rules for the development of Artificial Intelligence (AI) systems in the country.

PL No. 21/2020 aims at creating a legal framework for AI in Brazil through “principles, rights and duties” for the use of this technology in national territory. Note the importance of the definitions provided in the document, as these are the elements that in practice determine the applicability of rules or regulations to specific economic activities. Also relevant is the proposal for an “impact report” that the actors involved with AI would make available, dealing, in summary, with risk mitigation, management and protection mechanisms for users and legal and regulatory requirements.

Other relevant elements should be stressed in PL 21/2020, such as, for example, the scope of the definitions of “AI agents” and “stakeholders”; the concern to seek a balance between innovation, competitiveness, economic growth, development and protection of fundamental rights; principles (broad benefit to human beings and society, centrality in the human being, transparency, explainability, security, responsibility and accountability) that reflect the main international proposals in this space; and intersection between PL and LGPD.

PL 240/2020 currently attached to PL 21/2020, of a more generic and axiological character, proposes the setting of principles, guidelines and technical requirements for AI systems, leaving the Executive Branch in charge of creating a national policy that regulate this technology.

In addition, in the Chamber of Deputies this year, two requests (REQ 99 and 434) were submitted, which request the registration of the “Mixed Parliamentary Front of Artificial Intelligence”, installed in February 2020 and the creation of a Special Commission to propose the framework regulatory framework, under the justification of debating the guiding principles for the application of technologies derived from AI in society, in order to develop guidelines that encourage the development and use of these innovations in the private and public sectors.

In the Federal Senate, two bills presented in 2019, with a determination in February 2020 to hold a public hearing at the “Commission for Science, Technology, Innovation, Communication and Informatics” to instruct the matter, have the objects related to the setting of guidelines for the use of AI.

PL 5051/2019 aims to establish “the principles for the use of AI in Brazil”. The external writing aspects related to the centrality of AI systems in the human being, jobs, performance of Public Administration, human supervision and responsibility regime. To this end, it stipulates four groups of principles linked to fundamental rights, human rights, data protection and privacy and requirements that minimize the risks associated with the use of this technology.

PL nº 5691/2019 aims to institute “the National AI Policy” focused on stimulating innovation, economic growth and the organization of a favorable environment for this technology. Thus, it stipulates principles complemented by conduct parameters as mandatory requirements for AI systems, which, together, show the primary concern for human autonomy, risk mitigation, transparency, stakeholder participation and protection of human rights and democratic values.

In creating legal standards that determine rules applicable to the lifecycles of AI systems, it is necessary to be guided by the flexibility and importance of regulations. Now, this type of technology advances rapidly and exponentially, involving, in many cases, different jurisdictions and actors. Therefore, it is essential to establish general rules, accompanied by specific regulations, followed by a wide debate, in which companies, technical and scientific authorities and civil society participate.

In this context, the issue of the liability regime stands out, considering the multiplicity of actors involved, which presupposes that the rules and norms in this area consider the position of these actors to respond to risks and developments from the development to the availability of products and services, their use and decisions taken autonomously.

Another important aspect concerns human intervention. It is understood that the discussion needs to be guided by reflection on instruments that achieve this objective, such as, for example, the link between the practical effects of AI and human validation, the possibility of auditing and subsequent checking by human persons, or even deactivating the application during use.

By: Wilson Sales Belchior