European parliamentarians have approved the Artificial Intelligence Act. This was reported by the press service of the European Parliament. 523 people voted “for”, 46 “against”, 49 abstained.

The law is intended to make some uses of technology illegal and require AI solution providers to provide transparency.

The legal text of the document is still awaiting final approval. The provisions of the law will come into force in stages: in the first 6 months, EU countries will block the operation of prohibited AI systems; 12 months are given to ensure compliance with regulations regarding “general purpose AI systems” such as chatbots; and will take up to 36 months to approve enforcement of AI systems designated as posing a “high-level risk.” The rules of law are divided according to the level of risk that AI systems pose to society - “the higher the risk, the stricter the rules.”

The law prohibits the use of systems such as social rating platforms, emotion recognition platforms in workplaces and educational institutions, and systems designed to influence citizen behavior or exploit vulnerabilities. “High-risk” AI systems are used in critical infrastructure, education and training, law enforcement, and systems that could influence democratic processes, including elections.

Some EU countries have expressed concerns that overly strict regulations will make the region less attractive to the artificial intelligence market. Germany, Italy and France have advocated easing restrictions on general-purpose AI.