The European Parliament has given this Wednesday the ‘green light’ to the first rules of the European Union (EU) for Artificial Intelligence (AI), which will now negotiate with the Council, to prohibit biometric surveillance and impose transparency in systems such as ChatGPT.
At the plenary session of the European Parliament, in the French city of Strasbourg, the MEPs approved with 499 votes in favour, 28 against and 93 abstentions their negotiating position to start talks with the EU Member States (represented in the Council) on the final form of the AI law, which the Commission hopes will come to fruition later this year.
“The standards aim to promote the adoption of reliable and human-centered AI and protect health, safety, fundamental rights and democracy from its harmful effects,” the European assembly stressed in a statement.
Specifically, in this position now adopted by the MEPs, it is defined that the new regulations provide for a total prohibition of AI for biometric surveillance, emotion recognition and preventive surveillance, and impose that the systems that generate this technology such as ChatGPT indicate in a transparent way that the contents have been generated by AI and although the programs used to influence voters in elections are considered high risk.
In April 2021, the European Commission presented a proposal to regulate AI systems, the first legislation at EU level and which aims to safeguard the fundamental values and rights of the EU and the safety of users, forcing systems considered high risk to comply with mandatory requirements related to their reliability.
Since then, the document has been discussed among the co-legislators and, in an interview with the Lusa agency and other European media in Brussels at the end of May, the executive vice-president of the European Commission responsible for A Europe Prepared for the Digital Era and for Competition, Margrethe Vestager, said she expects an agreement this year in the EU on the first law on AI.
“We hope to have the first three-way dialogue [discussão entre colegisladores] on the AI Law in Europe before the summer, which would mean that if we work hard for it, we can have a result by the end of the year,” said Margrethe Vestager, although she admitted that the new rules will only come into force in 2025, given the necessary adaptation time, not following the speed of technology.
This will therefore be the first regulation aimed at AI, despite the fact that the creators and developers of this technology are already subject to European legislation on fundamental rights, consumer protection and standards on product and safety safety. responsibility.
Additional requirements are expected to be introduced to deal with risks, such as the existence of human supervision or the obligation to provide clear information on the capabilities and limitations of artificial intelligence.
AI has been increasingly used in areas such as entertainment (content personalization), ‘online’ commerce (predicting consumer tastes), household appliances (intelligent programming) and electronic equipment (use of virtual assistants such as Siri or Alexa, among others) .
The European Commission has been trying to strengthen cooperation between Member States on AI, but there is still no common legal framework, so the aim is to move from a voluntary approach to the regulatory sphere.
Source: TSF
