Anthropic PBC has disclosed that its Claude chatbot now leads over a quarter of the company's AI research and development efforts, offering the clearest evidence to date that AI can accelerate the creation of future models. According to a report released by Anthropic on Thursday, Claude "dominates" 26% of its R&D, up from virtually zero at the start of this year. The AI developer also stated that Claude assists employees in roughly 90% of their work tasks.

These figures are part of a broader report from Anthropic aimed at helping the public monitor AI's trajectory, amid growing concerns that the technology could rapidly enhance its own capabilities and slip beyond human oversight. In the report, the company reiterated its plan to bring in third-party evaluators from multiple institutions, granting them the same level of access to internal processes, systems, and data as its own employees.

"AI systems are becoming more powerful at an exponential rate and have begun automating more of the processes that build them. As the world considers slowing down the development of frontier AI, the public needs more information," the company wrote. Anthropic is one of several AI firms focusing on "recursive self-improvement," a concept where AI systems can iteratively advance themselves with minimal or no human intervention. While some view this as the future of AI development, its realization has also sparked concerns about potential safety risks.