On September 28, 2026, the Cambridge Programme on AI Science & Policy released a 22-author paper arguing that automating AI research could sharply accelerate progress.
The authors, including Geoffrey Hinton, Yoshua Bengio, and Jack Clark, say AI systems are on track to automate most AI R&D work within a few years, creating a loop where stronger systems help build the next generation faster. To support that case, the paper cites Anthropic data showing AI systems’ share of approved code rising from low single digits to over 80% between January 2025 and May 2026.
It also points to internal R&D work completed with only high-level supervision rising from 1% to 26% between March and August 2026. In a more tentative section, the authors say some extrapolations suggest months-long AI R&D projects could be automated by mid-2028. The paper’s policy ask is for governments to get better visibility into AI R&D automation and develop ways to steer or constrain the process.