EU AI Act Prohibited Practices: Already Banned Since February 2, 2025
The EU AI Act's most restrictive provisions — the Article 5 prohibited practices — took effect on February 2, 2025. These are not upcoming deadlines. They are ALREADY illegal. If your organization uses any of these AI practices involving EU residents, you are currently non-compliant.
Timeline
EU AI Act entered into force
August 1, 2024
Article 5 prohibited practices banned — NOW IN FORCE
February 2, 2025
TODAY — enforcement ongoing
April 15, 2026
Full enforcement for all other obligations begins
August 2, 2026
Violations of Article 5 carry the highest fines in the entire EU AI Act: up to €35,000,000 or 7% of global annual turnover, whichever is higher. These fines are larger than GDPR because prohibited practices represent the most severe risks.
Any organization deploying or using AI systems involving EU residents in any of the prohibited categories below.
What to do right now
Audit: Do you use any AI-driven social scoring system? → Stop immediately if targeting EU residents.
Audit: Do you use subliminal manipulation techniques in AI? → Stop immediately.
Audit: Do you use AI exploiting vulnerabilities of vulnerable groups? → Stop immediately.
Audit: Do you use real-time remote biometric identification in public spaces? → Prohibited except in very narrow law enforcement exceptions.
Audit: Do you infer political, religious, or other sensitive attributes from biometric data? → Prohibited.
Document your assessment confirming none of your AI systems fall into prohibited categories.