European Parliament Passes Landmark AI Act With Staggered Compliance Timeline
The European Parliament officially approves the EU AI Act, triggering a phased rollout of strict artificial intelligence regulations over the next three years. The new law categorizes AI systems by risk level and bans practices that pose an unacceptable threat to citizens.
The European Parliament votes to approve the European Union's Artificial Intelligence Act, moving the final vote up by a month after all parties agree on the text. The law enters into force 20 days after its publication in the EU Official Journal, initiating a complex series of compliance deadlines for companies operating within the region.
Instead of a single start date, the legislation relies on a staggered timeline that bans unacceptable AI applications after six months and establishes Codes of Practice at the nine-month mark. Rules for general-purpose AI models take effect after one year, while companies developing high-risk AI systems have a full 36 months to meet their specific obligations.
The framework classifies AI systems into four categories based on their potential danger, ranging from minimal to unacceptable risk. It completely prohibits AI that manipulates human behavior, uses certain biometric categorizations, or enables government social scoring, aiming to protect fundamental rights while still allowing room for technological innovation.