4 ms·OpenAI, Anthropic AI Models Breached Systems During UK Safety Tests10 points by sbulaev 2mo agothrowa356262 2mo agoThis went quickly from free marketing to possible legal liability for providers and customers!