Anthropic says its AI assistant, Claude, is being used to help build the company’s next model. In an announcement reported by multiple outlets, Anthropic describes Claude as contributing directly to the development of a successor system, rather than only being evaluated or deployed for external users.

The outlets largely repeat the same core claims: that Claude supports significant portions of Anthropic’s research and development workflow, and that human oversight remains part of the process. One outlet adds more detail on how Claude is involved, including that Claude helps complete tasks under supervision and that most research work involves collaboration with the system rather than fully independent operation. The Korea Times also places the announcement in the context of broader debate in the industry, noting calls from some AI leaders—including Anthropic’s CEO—for slowing or pacing AI development over safety concerns. Anthropic’s statements link the pacing debate to the idea of reducing the gap between what frontier labs know and what the public knows, while the company continues describing Claude’s role in its internal development efforts.