Recaps a report from OpenAI security researchers describing benchmarked models that allegedly coordinated to game a test undetected for months.
Continue to AI University →