A co-founder of Anthropic has projected that continued advances in artificial intelligence could lift U.S. gross domestic product by about 50% over four years.
Jack Clark, head of public benefit at Anthropic, made the forecast in a speech at the annual meetings of the International Monetary Fund and the World Bank in Bangkok on the 17th, Reuters reported.
Citing the company's internal analysis model, Clark said U.S. GDP, now around $30 trillion, could reach $44 trillion by 2030. "I'm well aware that this number sounds very strange," he said, explaining that AI is dramatically accelerating scientific discovery.
He added that the outlook rests on the assumption that the technology keeps advancing, and that if such economic expansion materializes, it would also generate new industries and inventions to drive it.
Clark described a future in which AI solves a range of problems in mathematics, engineering, biology and drug development, with human researchers assisted by hundreds or thousands of AI helpers. While acknowledging the reality of AI-driven layoffs and job losses, he said AI is boosting employment in other areas such as scientific research and has the potential to create jobs in entirely new industries.
He cautioned, however, that as AI advances more quickly, evidence is mounting of a "singularity" at which the technology outpaces human control. "By being able to use AI for science, we may be exposed to the misuse of science, such as biological and cyber weapons, or to things we have not yet imagined, and we will have many choices to make," Clark said.
Although there have been a string of recent incidents in which AI agents hacked outside organizations without human direction, Clark said AI "is still under human control," while stressing that only a few years remain to put safeguards in place to prevent AI from escaping that control.
Clark, a former Bloomberg reporter, joined OpenAI in 2016 and left the company in 2021 with Dario Amodei and others to found Anthropic.