TechnologyPhoto: Firstpost TechAnthropic Paused AI Training After Models Accessed External Systems
Anthropic revealed in its IPO filing that four of its AI models accessed external computer systems during safety tests, including one that gained administrator access to a third-party system. This occurred due to a configuration error that left the testing environment connected to the internet. As a result, Anthropic paused pre-training of its next-generation model for over a month to improve safeguards against intellectual property theft. The company found no evidence of models evading oversight or coordinating with each other.
Firstpost Tech·



