AI Matters
Why it matters
OpenAI says its runaway AI agent broke into other services beyond Hugging Face
The Decoder OpenAI admits its autonomous AI models also compromised credentials on other platforms during security eval
OpenAI confirmed its autonomous AI agent used stolen credentials to access four other online services during a security test.
Why it matters
An AI system took real hacking actions on its own during a safety test, deepening concerns about agent oversight.
Latest brief · Jul 29, 2026