The European Union's new AI Act took effect on August 1, 2024, marking a significant step in regulating artificial intelligence technologies.

Finalized in May 2024 after four years of negotiations, the Act aims to establish a global governance framework for AI, with member states required to designate national authorities by August 2, 2025.

The legislation adopts a risk-based approach, categorizing AI systems as low, medium, or high-risk, with stricter regulations for high-risk applications that must be safe, transparent, non-discriminatory, traceable, and environmentally friendly.

Generative AI, classified as General Purpose AI (GPAI), faces specific regulations, including a ban on certain high-risk applications, such as social scoring, six months after the Act's implementation.

AI systems classified as posing limited risk will have transparency obligations, such as informing users they are interacting with AI and labeling AI-generated content like deepfakes.

Bans on certain AI applications will be rare and specifically target systems that violate European values, such as citizen scoring systems or mass surveillance technologies.

By August 2025, most provisions of the law must be fully implemented, while high-risk systems will have an extended deadline of three years to comply.

Critics argue that the legislation could either stimulate AI development or hinder it, depending on national implementation and how companies adapt to the new rules.

The law's extraterritorial scope requires any company with EU operations to comply, regardless of its headquarters location, emphasizing the global impact of the legislation.

Over 700 companies have pledged to comply early with the AI Act through an AI Pact, indicating a proactive approach to the new regulatory environment.

Although the AI Act is now in force, many provisions will not be effective until 2026, allowing a transition period for existing generative AI systems to comply.