The report documents attempted misuse across cyberattacks, influence operations, surveillance, biology, and weapons development, all of which the company disrupted.
AI developers and security teams gain visibility into sophisticated attack vectors and misuse patterns targeting frontier AI models.