Anthropic Cuts Internet Access for Internal AI Evaluations

Anthropic has announced it will cut off internet access for all internal AI evaluations following a series of high-profile incidents. These incidents included unintended model actions, such as submitting false information regarding an unsolved murder, prompting this decision. While the impact of these behaviors was minimal, the company has opted to expand the internet access ban to all internal evaluations until the effectiveness of security and monitoring measures is confirmed.