Anthropic’s Jack Clark predicts AI will help win a Nobel within a year

TL;DR:

• Anthropic co-founder Jack Clark told an Oxford University lecture on Wednesday that AI will work with humans to make a Nobel prize-winning discovery within 12 months, and that bipedal robots will assist tradespeople within two years.
• Clark predicted that companies run solely by AIs would be generating millions of dollars in revenue within 18 months and that by the end of 2028 AI systems would be able to design their own successors.
• He maintained that there remained plausible scenarios in which the technology had “a non-zero chance of killing everyone on the planet” and said it was “important to clearly state that that risk hasn’t gone away”.

Clark’s lecture lands in a week thick with corroborating signal. OpenAI announced the same week that an internal model had autonomously disproved Erdős’s unit-distance conjecture, with Fields medalist Tim Gowers calling the proof a “milestone in AI mathematics”. Anthropic told investors it is closing in on its first quarterly operating profit. The £670bn ($900bn) company’s most senior public voice is offering this set of predictions from a podium at Oxford in front of an Institute for Ethics in AI audience.

The most-quoted UK AI lecture of 2026 so far

The Guardian’s coverage emphasises a duality that Clark has been running for years: the technology is moving faster than people realise, and the risks are real. “If we stand by and let synthetic intelligence multiply, then we’ll eventually be forced into reactivity,” he said, comparing the failure to prepare for AI to the failure to prepare for pandemics including Covid. He acknowledged Anthropic has been accused by the Trump White House and other AI accelerationists of “fear-mongering” to entrench its competitive position; he also conceded that international competition between actors and countries was “drowning out the larger existential-to-the-species aspects” of the build-out.

Professor Edward Harcourt, director of the Institute for Ethics in AI which co-hosted the lecture, separately warned of “cognitive atrophy” as humans hand more tasks to AI — advocating “Socratic” AI models that prompt humans to do more thinking. The two sets of remarks together capture the two poles of the UK academic AI debate: capability acceleration and human capacity preservation. Clark’s most conservative prediction — that “vast swathes of the economy and society will go through profound changes”, potentially including a machine economy decoupling from the human economy — admitted some of what he was saying sounded “crazy”. That admission itself is news; the head of Anthropic’s policy effort is on record saying his most conservative forecast sounds crazy.

Looking forward

Watch for whether Clark’s 12-month Nobel timeline survives the next AI-research announcement cycle, given that OpenAI’s Erdős result this week is now part of the publicly verifiable case. UK universities and research bodies — the Royal Society, the EPSRC, the Royal Academy of Engineering — will need to respond to the implied research-acceleration timeline. For UK businesses, the value in Clark’s framing is the explicit prediction calendar. If AI-run companies generating millions of dollars in revenue by late 2027 turns out right, the operating-model implications for UK SMEs and corporates will arrive faster than current strategic plans assume.

Currency conversion uses £1 = $1.27 at time of writing.