Draft EU AI Regulation Would Ban Social Scoring, Limit ‘High-Risk’ Uses
April 14, 2021
A leaked draft of the European Commission's proposed AI regulation includes a ban on social scoring and strict limits on mass surveillance and other "high-risk" uses of AI. The draft regulation would generally prohibit AI which "manipulates human behaviour, opinions or decisions" to a person's detriment or which "exploits information or prediction about a person or group of persons in order to target their vulnerabilities[.]" The draft also requires notice to individuals when they interact with AI, prior authorization for the use of remote biometric identification tools (including facial recognition), and data impact assessments for "high-risk" systems. The draft is broadly worded and subject to exceptions—including exemptions for "investigating serious crime and terrorism"—but would impose a penalty of up to 4% of annual revenue on companies that violate the regulation. The official release of the proposed regulation is expected on April 21. EPIC has called for prohibitions on secret scoring, mass surveillance, and facial recognition.