The European Commission officially proposed the Artificial Intelligence Act on the 21st of April 2021. This proposal initiated a legislative process that lasted nearly three years before reaching an agreement.
Non-compliance with prohibitions in Article 5 subjects offenders to administrative fines up to thirty-five million euros or seven percent of total worldwide annual turnover whichever amount is higher. Other operator obligations carry sanctions of fifteen million euros or three percent of worldwide annual turnover.
Applications classified as unacceptable risks are banned outright unless specific exemptions apply including systems that manipulate human behavior or use real-time remote biometric identification in public spaces. Social scoring systems that rank individuals based on personal characteristics also fall under this prohibition.
General-purpose AI foundation models face unique transparency requirements and must publish training data summaries if their weights and design are made open source. High-impact models posing systemic risks require extra evaluations including adversarial testing and must report serious incidents while ensuring adequate cybersecurity measures.
The regulation entered into force on the 1st of August 2024 twenty days after being published in the Official Journal on the 12th of July 2024. Provisions come into operation gradually over six to thirty-six months with bans taking effect after six months and high-risk system obligations having a thirty-six-month implementation window.