AI Ethics Assessor
Policy Update

Outcomes of the Final EU AI Act Negotiations

January 2024

Negotiation Outcomes

Prohibited Systems

The latest round of negotiations has led to a political agreement to prohibit several controversial technologies:

  • Social Scoring: Systems that exploit vulnerabilities or use manipulative techniques.
  • Facial Scraping: Creating databases by scraping images from the internet (e.g., Clearview AI).
  • Emotion Recognition: Prohibited in workplaces and education, with narrow exceptions.
  • Predictive Policing: Individual risk assessments to infer criminal likelihood.
  • Biometric Categorization: Based on sensitive traits like race or religious beliefs.

Severe Penalties for Non-Compliance

The repercussions for violating these rules are significant. Organizations could face:

  • Fines up to 35 million euros or 6.5% of global turnover for using banned applications.
  • Fines up to 15 million euros or 3% for high-risk system violations.
  • Potential market bans for the most serious infractions.

Impact on Open Source

The Act aims to support innovation by exempting free and open-source components from most regulations, unless they are part of a high-risk system. It recognizes the economic value of the open-source ecosystem, estimated at up to EUR 95 billion for the EU GDP.