The Brazilian Senate has released a 900+ page report & draft AI Law, drawing inspiration from earlier proposals & research into how OECD countries regulate AI. Here are some aspects I read in the draft AI law...
The Draft AI Law in Brazil requires developers, implementers and users of AI to adhere to principles of good faith, self-determination, transparency, non-discrimination, legality, reliability and proportionality. #AI #brazil
It defines an AI system as a computational system with varying degrees of autonomy, designed to achieve objectives using ML/logic & knowledge representation to produce predictions, recommendations, or decisions that can influence the virtual/real environment.
Prior to deploying an #AI system, providers and users must conduct and document a risk assessment. #RiskAssessment
The draft classifies health applications, autonomous vehicles, as “high-risk”, and the competent authority is responsible for updating this list. Among others are: securing operation of critical infrastructure; education; recruiting;
credit scoring; use of autonomous vehicles; and biometric identification. The competent authority is responsible for creating and maintaining a public database of high-risk AI systems, which will include risk assessments.
These assessments should be protected by applicable intellectual property and trade secret laws.
Brazil's draft prohibits the use of AI systems that deploy subliminal techniques or exploit vulnerabilities of specific groups, as well as social scoring & biometric identification systems in publicly accessible spaces,unless there's specific law or court order that authorizes
Rights: Individuals affected by AI systems have the right to information on their interactions with the system; right to an explanation about a decision, recommendation, or prediction made by an AI system within 15 days of the request;
the right to challenge decisions; Right to human intervention in decisions made solely by AI systems; Right to non-discrimination and the correction of discriminatory bias; Right to privacy and the protection of personal data, accordingly to Brazilian General Data Protection Law
Providers and users of all AI systems must establish governance structures and internal processes to ensure security and protect individuals' rights, while providers and users of "high-risk" AI systems must implement additional measures such as algorithmic impact assessments
and teams to ensure diverse viewpoints. Codes of conduct and governance may also be drawn up to support the implementation of the law.
civil liability: Providers and users of AI systems are liable for any damages caused by their AI system, with providers and users of "high-risk" AI systems being strictly liable for their participation in the damage and their fault being presumed.
The draft states that the automated use of existing works (extraction, reproduction, storage and transformation) by AI systems for research organizations, journalists, museums, archives and libraries, may not constitute a copyright infringement under certain scenarios.
The Brazilian Government must appoint a regulatory authority to enforce the law, which carries potential fines of up to 50 million Reais (approximately 9 million Euros) or 2% of a company’s turnover for any violations.
The 900+ pages are a starting point for the senate to creating and approving a bill that will be discussed in the Chamber of Deputies...
For those who are asking here is the link to the draft www6g.senado.leg.br/sdleg-getter/d…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
