EU AI Act — Risk Classification

The EU AI Act is the world's first comprehensive AI regulation. It classifies AI systems by risk tier and defines the corresponding obligations for providers and users.

European Union EU AI Act

What is the EU AI Act?

The AI Act (Regulation (EU) 2024/1689) entered into force on 1 August 2024 and is being phased in until 2027. It follows a risk-based approach: the higher the risk of an AI system, the stricter the requirements.

The 4 risk tiers

  • Unacceptable (prohibited) — Social scoring, manipulation, real-time biometric surveillance (with exceptions). These AI systems are banned.
  • High risk — Critical infrastructure, education, employment, law enforcement, migration. Strict requirements on transparency, data sets, and human oversight.
  • Limited risk — Chatbots, deepfakes, emotion recognition. Transparency obligations: users must know they are interacting with AI.
  • Minimal risk — Spam filters, recommendation systems, video game AI. No specific obligations, but codes of conduct are recommended.

Risk tier quick check

Answer 6 questions about your AI system to determine the likely risk tier.

e.g. subliminal manipulation, exploiting age, disability, or social situation

e.g. trustworthiness based on social behavior, personality traits

Critical infrastructure, education/training, employment/HR, credit scoring, law enforcement, migration, justice

e.g. applicant screening, creditworthiness checks, social benefits, diagnoses

Or does it generate synthetic content (text, image, audio, video)?

Real-time face recognition in public places

Inspiriert von European Union — EU AI Act

Weiterlesen

Trivia

  • The EU AI Act is the world's first comprehensive law regulating artificial intelligence.
  • Negotiations lasted over 3 years (April 2021 to March 2024).
  • General-purpose AI (like GPT-4, Claude) has its own rules: transparency obligations and copyright compliance.
  • Violations of the AI Act can be fined up to 35 million euros or 7% of global annual turnover.
  • The prohibitions (unacceptable AI) have been in effect since February 2025; the remaining rules follow in stages through August 2027.