GMAsia
    AI News·17 Sept 2026·via Engadget

    Anthropic says Claude 'leads' 26 percent of its AI R&D work

    Anthropic reports its AI chatbot Claude now leads 26 percent of its AI research and development work. This means Claude completes most of a task end-to-end under human supervision. The company also states AI handles large chunks of work under close human direction for over 90 percent of its research. These metrics are part of Anthropic's new proposed standards for measuring AI development pace and oversight.

    Nexa's Summary

    Anthropic's claim that Claude "leads" 26 percent of its AI R&D work is a significant data point. It shows a frontier model directly influencing a substantial portion of its own creation. This 26 percent figure, coupled with AI touching over 90 percent of research work, reflects increasing AI integration in development. It is a concrete measure of how quickly AI is becoming a co-developer.

    For Asia's AI developers, this creates a growing competitive pressure. Companies like Alibaba Cloud and Tencent AI Lab must now consider similar internal AI-led R&D metrics. The ability for a model like Claude to accelerate development internally could create a speed advantage. Asian firms need to assess if their own models can achieve comparable automation levels to remain competitive.

    The thing to watch is whether other frontier AI labs adopt Anthropic's proposed measurement standards. If companies like Baidu or SenseTime begin publishing similar metrics, it would increase transparency. This would offer a clearer view of the actual pace of AI development across the industry.

    #ai
    Go deeper
    Original reporting by EngadgetWe don't republish, read the full story →

    Related reading

    6 stories