Last May, AI Act was adopted by the member states of the European Union. This ambitious text aims above all to protect the rights of citizens and limit the abusive uses of AI. While the first provisions of this legislation came into force on Sunday, February 2, here is everything you need to know about this new device.
What measures are already applied?
In detail, the text now prohibits all private or public social rating systems, as can be observed in China. Predictive police systems that aim to estimate whether a person is more likely to commit offenses are also prohibited.
The same goes for applications of “Recognition of emotions” at work and at school, details The world. The identification of citizens by real -time facial recognition in public space and the biometric categorization of people to identify “Their race, their political, religious opinions, their sexual orientation or their union membership”are finally prohibited.
What calendar for other provisions?
Our colleagues specify that from August 1, 2025, the “general use” AI models, in other words text and image generators such as Chatgpt or Gemini will have to show transparency in their technical documentation and on their training data. Cyber risks safety audits will also have to be carried out.
Much later, in 2027, AIs judged at high risk, such as those operating in water, road electricity infrastructure, justice, police or education will have to carry out assessments concerning risks of errors and bias.
Who will control the application of AI Act?
To ensure compliance with the AI Act rules, an AI office will see the light of day. He will in concert with the National Data Protection Authorities, such as the National Commission for Data Protection (CNIL) in France. If they contravene the rules, the companies concerned are led to savory fines of up to 7% of their global turnover.
Why is this text debate?
From its adoption, AI Act suffered strong criticism from actors in the AI sector. After The worldit is notably the dissemination of a sufficiently detailed summary of the content used to form the large models that poses a problem.
Right holders are delighted with this principle, as they hope to receive compensation if their productions have been used. But the tech giants on the contrary oppose the argument of business secrecy to reject this obligation.
Will Act Act brake EU’s competitiveness?
Some observers have argued that the EU could shoot itself in the foot by imposing binding standards on AI companies. On the French tech side, we also plague against the risk of the regulatory millefeuille which would brake European innovation.
Quoted by Le FigaroKirsten Rulf, Director associated with the Boston Consulting Group (BCG), co-author and negotiator of the EU Ai Act, on the contrary defends the approach of the text:
AI Act gives an action plan to improve the quality of these systems and therefore, on the contrary, promote innovation. We are on technologies still very unstable, which can make mistakes and jeopardize a company if the risks are not well controlled.
Same story on the side of Mahasti Razavi, associated with August Debouzy specializing in Tech issues. Questioned by The echoesthe expert precise: “To schematize, this text is there to create a framework of confidence”.
In the coming months, AI Act promises to be talked about a lot. And he may well be part of the discussions between the EU and the government of Donald Trump, which should support the grievances of the tech giants.
