The Washington Post reported that European Union legislators reached a historic deal on a broad-ranging AI safety development bill after a marathon 72-hour debate. The proposed regulations would impact the development and distribution of machine learning models within the trade bloc, affecting applications in areas such as education and healthcare. AI development would be categorized based on societal risk, with banned uses including those that circumvent the user’s will, target protected groups, or provide real-time biometric tracking. High risk uses would include applications in critical infrastructure, education, legal/judicial matters, and employee hiring. Dr. Brandie Nonnecke, Director of the CITRIS Policy Lab at UC Berkeley, commented that the European Commission’s approach is similar to Canada’s proposed AI regulatory framework, addressing the problem from a risk-based perspective.
EU Reaches Historic Regulatory Accord on AI Development
