Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead

Anthropic announced it has turned off live internet access for all its internal evaluations until further notice. The step was taken because the company cannot reliably control its AI agents when they are connected to the web, raising safety concerns.

Summary written by Kernelia from the original article by TechCrunch AI. The story and its rights belong to its author.