A self-driving car runs over a pedestrian. Who is to blame? This is the most classic example in the discussion about the technological advancement of artificial intelligence (AI). With the Chamber of Commerce approving a draft legal framework for artificial intelligence, Brazil is about to take its first steps toward concrete regulation of the topic. But experts say that contrary to expectations, the proposal does not provide many answers or guidance.
Pixar
AI is increasingly common on Pixabay marketplace
Artificial intelligence is the programming of systems, applications, and machines to "imitate" human abilities. In other words, the robot can interpret the external environment and "learn" from it to perform tasks and make decisions. Back to the example of the AI-powered vehicle running over a pedestrian: What if the collision happened because the “driver” decided it would be more fun not to let his car hit a Ferrari?
In addition to self-driving cars, these technologies are also used in smart home appliances, virtual assistants, and even the recruitment process of job candidates-there are already in-depth discussions about "algorithmic discrimination".
The PL approved by delegates commits to “establishing foundations, principles Chinese Overseas America Number Data and guidelines for the development and application of artificial intelligence” in the country. In fact, the text provides standards for action by federal entities and a list of concepts to be followed. However, experts consulted by ConJur believe that these regulations are not solid enough to guide the use of these technologies.
The need for regulation
Companies working in artificial intelligence believe there is no need to regulate the use of these technologies as this could limit their potential. However, Patricia Peck, partner at Peck Advogados and senior advisor to the National Data Protection Commission (CNPD), stressed that the legislation cannot be viewed solely from a private perspective. He noted that “artificial intelligence is also used in the public sector”.
“It is a mistake to think that regulation always limits potential. It can help you,” says Eduardo Magrani, partner at Demarest and president of the National Institute for Data Protection (INPD).
While recognizing the importance of regulating AI, Peck noted that very superficial legislation could be more harmful than a lack of standards. The lawyer said that the legal framework proposal still needs to be supplemented.
Summary:
"The wording of the PL is still very principled and in its infancy. It needs to be strengthened and 'supplemented' to become an enforceable and effective law that can guide the market," Peck said.
Cleia Viana/House of Representatives
The bill was approved by the House of Representatives in late September
Cleia Viana/House of Representatives
Magrani said the project was approved "without vigorous public debate" and therefore became superficial and simplistic. For lawyers, the drafting of PL even lacks principles. The concept of insertion remains "misguided and often insufficient."
Peck understands that there is a positive side to the concept and principles being clearly defined. The biggest problem is the lack of other minimum provisions that the law should have: liability, penalties, inspections, updated models (such as self-regulation and codes of conduct) and the obligation to promote educational activities on the subject.
The same criticism of a text based solely on ethical standards led a group of experts linked to the LGPD Institute to submit a proposal to the Chamber of Commerce to improve the PL, inserting more realistic predictions.