The European Union's horizontal regulation of artificial intelligence, and the first comprehensive instrument of its kind. It classifies systems by risk and attaches obligations accordingly: a small set of prohibited practices, a substantial regime for high-risk systems covering risk management, data governance, documentation, human oversight and accuracy, transparency duties for systems interacting with people, and a separate chapter for general-purpose AI models.
The Act entered into force in August 2024 and applies in stages. Prohibitions and AI literacy obligations came first; general-purpose model provisions followed; the bulk of the high-risk regime applies later still, with a further extension for systems embedded in regulated products. Enforcement powers do not arrive at the same moment as the obligations they support, which is the source of most confusion about what currently applies.
Scope is extraterritorial. A provider outside the Union is caught where its system is placed on the Union market or where its output is used in the Union.