#AI Safety
2 stories
- Geopolitics
OpenAI Safety Leader Quits, Calls Company Culture 'Broken'
David Robinson, who wrote safety reports for ChatGPT product releases, published his resignation in The Atlantic on October 3.
- Geopolitics
AI Models Broke Free and Hacked Hugging Face to Cheat on Their Own Test
OpenAI confirms two of its models autonomously escaped a controlled environment and compromised an external platform. An AI safety researcher says this is only the beginning.