Brazilian lawmakers are seeking a more cautious approach to AI, in an attempt to align with the OECD
A group of experts to draft a new bill: The Brazilian Senate is establishing an ‘experts commission’ on artificial intelligence, to help them establish rules, principles, and guidelines to developing AI in Brazil. The commission’s main objective is to help the Senate draft a new bill to replace those proposed between 2019 and 2021, which were criticised by some stakeholders (including the Brazilian data protection authority) for not being subject to sufficient participation and scrutiny. The commission is made up of notable jurists, in recognition of the “technical, legal and moral” complexity involved in the use of AI. It will have 120 days to complete its work, during which it will receive input from civil society and hold seminars and public hearings.
A move in keeping with Brazil’s ambition to join the OECD: The criticism voiced by the leader of the National Data Protection Authority (ANPD) just days earlier may have played a role in the creation of the Commission. The ANPD noted that Brazil was rushing to legislate without thinking through the implications of regulating AI, and called for further study and further hearings. However, one of the main drivers behind the Senate’s decision is likely to be Brazil’s adherence to the OECD’s AI principles, at a time when Brazil is seeking accession to the organisation.
Taking a leaf out of the EU’s book: In its statement, the Senate also notes that the commission should take into account international examples of guidelines for the ethical use of AI, such as that of the European Union. The EU drafted its AI regulation following a lengthy consultation process with industry and experts in AI (expert group, criticism from France, etc.) which led to the ‘Guidelines on Trustworthy AI’ in 2019, followed by the AI white paper in 2020 setting out the four key policy objectives for it. This approach led to a detailed proposal based on different levels of risk – although some countries in the EU, such as France, are pushing for a more pro-innovation framework and criticised the EU proposal for being based solely on risk. AI frameworks present policymakers with a hard balance to strike, since overly prescriptive rules are likely to stifle innovation. At the same time, the concerns around possible serious harms to individuals and social groups are legitimate, and need to be heard.
