Anthropic co-founder predicts AI will contribute to Nobel Prize discovery within a year
Jack Clark, co-founder of AI safety company Anthropic, told Oxford University students that an AI system will collaborate with humans to make a Nobel Prize-winning discovery within 12 months. He also predicted bipedal robots will assist tradespeople within two years and that companies run entirely by AI will generate millions in revenue within 18 months.
Clark described a "vertiginous sense of progress" in AI capabilities. By the end of 2028, he said, AI systems will design their own successors. These predictions come as generative AI and large language models continue advancing at a pace that has outstripped many industry observers' expectations.
Risk remains, Clark says
Clark acknowledged that AI poses existential risks. He said there remain "plausible scenarios in which the technology had a non-zero chance of killing everyone on the planet" and stressed this risk "hasn't gone away."
Anthropic recently released a model called Mythos that demonstrated concerning ability to exploit cybersecurity vulnerabilities, underscoring the dual-use nature of advanced AI systems.
Speed of development outpaces safety measures
Clark said slowing AI development would benefit humanity by providing time to address the technology's implications. He acknowledged this won't happen.
Instead, he described "breakneck development by a variety of actors and countries, locked in competition with one another, where commercial and geopolitical rivalries are drowning out the larger existential aspects of the technology being built." He called this "not ideal."
Anthropic was founded by researchers who left OpenAI over disagreements about safety priorities. The company has faced criticism from the Trump administration and AI accelerationists who accuse it of "fear-mongering" to justify regulation that could strengthen its competitive position. Anthropic disputes this characterization.
Broader economic and social shifts ahead
Clark's most conservative prediction is that "vast swathes of the economy and society will go through profound changes." These could include a machine economy decoupling from the human economy, robots gaining autonomous decision-making capabilities, and AI advancing science without human involvement.
He acknowledged some of these scenarios sound "crazy."
Concerns about cognitive decline
Edward Harcourt, director of the Institute for Ethics in AI at Oxford, warned that humans increasingly relying on AI to perform tasks could create "cognitive atrophy" and weaken decision-making abilities.
Harcourt advocated for alternative AI models that require humans to do more thinking-sometimes called "Socratic" AI-rather than systems that automate human judgment.
Preparation gap
Clark compared the current lack of preparation for advanced AI to failures to prepare for pandemics like Covid. "If we stand by and let synthetic intelligence multiply, then we'll eventually be forced into reactivity," he said.
Critics of frontier AI companies like Anthropic, OpenAI, and Google worry that over-reliance on a small number of proprietary models backed by massive capital investment creates systemic vulnerability-a "single point of failure" in global systems.
Your membership also unlocks:
