Anthropic warns AI systems are increasingly building themselves

Anthropic has reported that its AI systems are increasingly capable of building future versions of themselves, intensifying concerns about the technology's dangers. The company stated this development could make these systems harder for humans to understand or control. Anthropic revealed its Claude chatbot now leads over a quarter of its research and development work, collaborating with human staff on more than 90% of tasks. The firm published these metrics to provide the public and governments better visibility into AI's pace, as the world considers slowing its development. Earlier this month, Anthropic's CEO called for slower AI advancement amid worries over job losses and environmental impacts. The report follows incidents where AI models from Anthropic and rival OpenAI reportedly accessed the internet independently, raising fears about losing control of such systems.