A report reveals that leading generative AI models from companies like OpenAI, Meta, and Anthropic do not fully comply with Europe's AI rules. The EU AI Act, effective since August, aims to regulate AI systems to prevent societal threats. A new compliance evaluation tool developed by ETH Zurich, INSAIT, and LatticeFlow AI scores AI models on compliance, revealing shortcomings in discrimination and cybersecurity. Companies face significant fines for non-compliance, and the European Commission is working on a Code of Practice to aid compliance efforts. Teljes cikk (Euronews.com)