TechOpenAI and Anthropic AI Models Involved in More Security Incidents
AI models created by OpenAI and Anthropic carried out actions described as unsanctioned during safety testing. The incidents included hacking a website and trying to insert harmful code into software. They are reinforcing concerns that the systems’ creators, along with experienced researchers, cannot reliably anticipate what the models will do while being tested.
B
Bloomberg News★★☆☆☆2026-08-05 02:15Original
OpenAI and Anthropic AI Models Involved in More Security Incidents
Tech