EU Set to Enforce AI Act on Social Scoring and Biometrics
European regulators plan to enforce AI Act on banned social scoring and biometric practices.
Why it matters: Non-compliance with the AI Act could lead to hefty fines; understanding its stipulations helps businesses avoid penalties.
- AI Act took effect on August 1, 2024.
- Enforcement on banned practices starts February 2, 2025.
- General-Purpose AI rules effective August 2, 2025.
- High-risk AI system regulations enforced by August 2, 2026.
The European Union's Artificial Intelligence Act, effective from August 1, 2024, is now entering phases of enforcement critical for businesses utilizing AI. Specifically, from February 2, 2025, enforcement will target AI applications that are considered of 'unacceptable risk.' This includes systems engaged in social scoring and real-time biometric surveillance.
The penalties for non-compliance with these provisions are severe, with fines reaching up to €35 million or 7% of the company's global turnover, whichever is higher. Such hefty fines necessitate vigilant compliance strategies for businesses operating within or into the EU.
From August 2, 2025, the Act imposes standards on General-Purpose AI systems, like ChatGPT, which require transparency and accountability measures. Analyst Enza Iannopollo from Forrester emphasizes the broad reach of these obligations due to the versatile application of these models across various sectors.
By August 2, 2026, full enforcement on high-risk AI systems is expected as per the AI Governance Desk's roadmap. Preparations are underway for investigations anticipated to begin in early 2026, setting a basis for cross-border regulatory efforts.
Fines for non-compliance with regulations on high-risk AI systems could reach up to €15 million or 3% of turnover. Understanding and preparing for these timelines is crucial for businesses as Member States are required to appoint national regulators by mid-2025 to ensure uniform application across the EU.
By the numbers:
- €35 million — Penalty cap for violating prohibited practices
- 7% — Global turnover fine for high-risk AI non-compliance
- August 1, 2024 — AI Act effective date
Yes, but: The enforcement could face challenges in international consistency and standardization due to varying local regulations.
What's next: Member States will appoint national regulators by mid-2025, with full enforcement expected by August 2026.