Hinton, Bengio and 20+ signatories warn automated AI R&D could trigger intelligence explosion

On September 27–28, a report titled "What if automating AI R&D triggers an intelligence explosion?" went viral. Co-signed by more than 20 authors from academia, civil society, and frontier AI companies—including Nobel laureates Geoffrey Hinton and Yoshua Bengio, as well as Jack Clark—the report assesses whether automating AI R&D could trigger an intelligence explosion and urges governments to act ahead of time.

Confirmed

• The report "What if automating AI R&D triggers an intelligence explosion?" was co-signed by over 20 authors spanning academia, civil society, and frontier AI companies (including Song Dawn, merettm, Jack Clark, Yoshua Bengio, and Geoffrey Hinton). It was covered by both The Guardian and Sky News.
• The report discloses internal Anthropic data: the share of AI-led tasks in its internal AI R&D rose from about 1% (as of February 2026) to about 26% (roughly a quarter) within roughly 5 months.
• The report treats automating AI R&D as one of the most consequential and dangerous technological trajectories today, arguing that it accelerates capability progress in unpredictable ways, could make AI systems more opaque to humans, and may erode human researchers' standing and bargaining power.

Why it matters

• AI helping build the next generation of AI has shifted from hypothesis to measurable reality: Anthropic's data offers the first concrete quantitative evidence from a frontier lab showing this process is accelerating R&D iteration within the lab.
• @victorexplore notes that while 26% does not mean full automation or "AI improving AI," it is already a substantive step in that direction.
• @SOhEigeartaigh cautions that external governance communities remain under-informed and under-prepared for this trend; the report aims to push governments to establish response mechanisms before intelligence explosion risks materialize.

2026-09-27 ~ 2026-09-28 · 7 related posts

• AI now leads a quarter of tasks at Anthropic, up from 1% in February — victor_explore · 2026-09-27
• [source] Report: Anthropic's supervised-only AI R&D share jumped from 1% to 26% in five months — S_OhEigeartaigh · 2026-09-28
• Researcher warns automating AI R&D is the most dangerous tech development ahead — S_OhEigeartaigh · 2026-09-28
• New paper with Bengio, Hinton: preliminary evidence says automating AI R&D could spark intelligence explosion — StephenLCasper · 2026-09-28
• [source] Hinton, Bengio and OpenAI/Anthropic execs warn of 'intelligence explosion' in 20-author report — nordicinst · 2026-09-28
• [source] Automation of AI R&D called the most dangerous tech on the horizon — new paper ships with Chinese translation — S_OhEigeartaigh · 2026-09-28
• What if automating AI R&D triggers an intelligence explosion? — pstAsiatech · 2026-09-28