What is the AI Act?

The EU AI Act imposes broad obligations on both developers and users of AI systems. Compliance requirements begin in August 2026.

What is the AI Act?

The AI Act is the world's first comprehensive legislative framework for artificial intelligence, aiming to harmonize rules on the use, marketing, and development of AI within the EU.

Who does the AI Act apply to?

• AI system providers
• AI system deployers
• Importers and distributors
• Third-country operators
• Product manufacturers

Risk levels of AI systems

1. Unacceptable risk AI systems – Prohibited in the EU. E.g., social scoring, manipulative systems, emotion recognition in schools and workplaces.
2. High-risk AI systems – Permitted but with strict requirements. E.g., traffic management, recruitment, credit scoring, robotic surgery.
3. Limited risk AI systems – Transparency requirements. E.g., ChatGPT, deepfake technologies.
4. Minimal risk AI systems – No specific requirements.

Timeline

• 1 August 2024 – AI Act entered into force
• 2 February 2025 – Prohibited practices provisions apply
• 2 August 2025 – General-purpose AI model rules apply
• 2 August 2026 – Full application of the regulation

Penalties

Up to EUR 30 million or 6% of global annual turnover.