From Engadget · Ian Carlos Campbell · · 1m 20s
Anthropic reveals Claude handles 26% of its AI research
The company proposes new transparency standards to measure how fast AI development is moving.
In brief
The company proposes new transparency standards to measure how fast AI development is moving. Anthropic is pushing for industry-wide metrics to track AI autonomy in research, showing Claude already handles a quarter of its R&D work with human oversight. Originally reported by Engadget.
Claude leads a quarter of Anthropic's R&D
Anthropic's AI chatbot Claude "leads" 26 percent of the company's AI research work, meaning it can complete most tasks end-to-end from a high-level prompt with human supervision.
“Claude "leads" 26 percent of its AI R&D work”
AI touches 90 percent of research
Claude does at least "large chunks of work under close human direction" on more than 90 percent of Anthropic's research, though the company says it's "not operating fully autonomously" for any measured subset.
“not operating fully autonomously for any measured subset of AI R&D work”
Three new measurement standards proposed
Anthropic created metrics to track AI-led R&D, oversight of AI agents, and compute devoted to research. The company used an automation rating scale from Epoch AI to plot Claude's capabilities since August 2025.
Transparency could prevent runaway AI
Anthropic hopes that clearer metrics and third-party validation will make it easier to respond to out-of-control AI development and give the public a better view of how fast the technology is advancing.
Industry commitment remains uncertain
While OpenAI has paid lip service to slowing AI development and Anthropic allows third-party review, President Trump has downplayed AI risks, leaving the future of regulation unclear.
Here's the gist
Anthropic is pushing for industry-wide metrics to track AI autonomy in research, showing Claude already handles a quarter of its R&D work with human oversight.
