RIO DE JANEIRO, BRAZIL – The government is progressing to govern the use of artificial intelligence in Brazil. The MCTIC (Ministry of Science, Technology, Information, and Communications) opened on Thursday, December 12th, a public consultation to define the so-called National Strategy for AI (Artificial Intelligence).
The plan is to debate ways to maximize the technology’s benefits and mitigate its negative impacts.
The measure comes after the country has acceded to the OECD recommendations in May, which established guidelines for governments to adopt systems that respect “human rights and democratic values”.
In the Senate, there is a bill advocating principles for a national AI policy.
The government has opened the discussion on legislation and ethical use of technology to the public until the end of January, considering areas such as workforce, research and development and implementation in the public and private sectors.
Following other countries that have adopted strategies, the government has brought this on the agenda after approving the National Plan for the Internet of Things and the General Law for Data Protection (in force next year), which mention automated systems.
“These strategies should guide the State on the development of public policy, not only in terms of regulation but also in terms of incentives and financing for the training of people who will need to be positioned in this market,” says Bruno Bioni, founder and professor of Data Privacy Brasil.
The consultation discusses what policies would be required for the re-qualification of professionals from sectors impacted by AI and what would be the role of the government in this regard.
From a governance perspective, it suggests the potential for compulsory prior impact assessment reports against discrimination (e.g. racial, gender or economic) in machine-based decision-making processes.
California, which has regulated facial recognition technology (it uses intelligent systems), has required public bodies to provide prior data protection reports before releasing such a technology. The documents need to be assessed by a committee to authorize it.
Canada, on the other hand, which has a declaration for the protection of rights in the use of learning systems, determines measures such as the clear and accessible availability of information on public systems that use machine learning for decision-making.
It also favors reports that reduce discrimination and the possibility of independent supervision through auditing systems.
The Canadian document calls for avoiding the use of “black box systems” that cannot be subject to accountability and transparency standards.