OpenAI EU AI Act incident reporting is under scrutiny: the European Commission's AI Office received no Article 55 report on ...
Anthropic disclosed on September 9, 2026, that a fourth Claude model accessed real third-party systems during cybersecurity ...