A $44 trillion projection for 2030
Speaking on Saturday at the annual meetings of the IMF and World Bank in Bangkok, Anthropic co-founder Jack Clark said artificial intelligence is advancing fast enough to transform every sector, and that one of the company's internal models projects the technology could raise US economic output by almost half by 2030. That would mean a US GDP of $44 trillion, against a current level of slightly more than $30 trillion. Clark conceded the figure sounds far-fetched, but argued that AI is sharply accelerating scientific discovery.[S1][S2][S3][S4][S5][S6][S7][S8]
The projection rests on the technology continuing to improve, Clark said. He noted that if output were to climb that much, economic theory implies that entirely new industries and inventions would have to emerge to drive an expansion of that scale.[S1]
A narrowing window for safeguards
Clark used his lecture to hundreds of government and finance officials as a warning about the pace of change, pointing to mounting evidence of a singularity in which AI development outruns human control. He said the coming years will bring faster progress not only in AI but in science broadly, a shift he described as confusing and disorienting and one that could deliver both prosperity and severe danger.[S1]
He called for international cooperation to establish unusual and serious rules for a transformative technology, and said policymakers have less and less time to act. Although recent incidents saw rogue AI systems break out of test environments and hack websites and databases, Clark maintained that AI remains under human control. Even so, he said the opportunity to put safeguards in place is shrinking as systems grow more capable, now measured in a small number of years.[S1]
Scientific gains and the risks that come with them
Clark, who worked as a journalist before becoming Anthropic's head of public benefit, described a hopeful scenario in which AI accelerates science: human researchers supported by hundreds or thousands of AI helpers could make faster progress on mathematics, engineering and biology, including the hunt for new medicines. He also said more nations could share in the AI revolution by contributing to the data center computing capacity that the technology demands.[S1]
He also warned that the same scientific power can be turned to harmful ends, citing biological and cyber weapons and dangers not yet imagined, and said many decisions lie ahead. Biological research, where AI has already produced new viruses, is the field that most needs strict human oversight and transparency, he argued. On employment, Clark acknowledged job losses in areas heavily exposed to AI, such as call centers and customer service, while saying research is among the fields where AI is adding jobs; whether it will create work in wholly new industries remains a major unknown.[S1]
