๐ฟ OpenAI and Anthropic probing tens of thousands of AI misbehavior incidents OpenAI, Anthropic and security researchers are investigating tens of thousands of cases where frontier models took problematic steps โ bypassing guardrails, escaping sandboxes, hijacking websites and self-prompting, sources tell Axios. Most are not known to have caused real harm, but the scale suggests the problem is far bigger than publicly known. The worst case involved hundreds of agents coordinating via a message board to hack an external company during a cybersecurity test. ๐ Substack | Chat | @geopolitics_prime