Clark delivered the warning on Saturday during the International Monetary Fund and World Bank annual meetings in Bangkok, Thailand, where he called for international cooperation to establish what he described as “unusual and serious rules” governing AI development and deployment.
He cited an internal forecast from Anthropic suggesting that, under certain conditions, artificial intelligence could help expand the United States economy to approximately $44 trillion by 2030, compared with its current size of just over $30 trillion.
Clark acknowledged that the projection might appear difficult to believe.
“Sounds very outlandish, I know,” he said, referring to the estimated economic expansion.
However, he argued that continued improvements in AI could dramatically accelerate scientific research, generate new inventions and create industries capable of supporting such substantial economic growth.
He explained that the forecast depended on the technology continuing to improve, adding that economic theory suggested such an expansion would require the emergence of new industries and innovations.
Clark’s remarks came as Anthropic prepared for an initial public offering reportedly expected in November, which could value the artificial intelligence company at as much as $2 trillion.
AI progress raises alarm over human control
Beyond the economic opportunities, Clark issued a warning about the speed of AI development, suggesting that the world could be approaching a point at which technological progress begins to outpace human oversight.
Addressing hundreds of government and financial officials, he said policymakers had a limited period in which to establish effective safeguards before AI capabilities advanced further.
“We should expect that in the coming years, the whole pace of progress of not just AI but science as a whole will speed up, and this will feel confusing and disorienting, and will bring with it great wealth and potentially great peril,” he said.
The concept of a technological “singularity”, in which AI development accelerates beyond human capacity to control or fully understand it, featured prominently in his warning.
AI was a major subject of discussion at the IMF and World Bank meetings, with officials weighing its potential to stimulate productivity and offset economic pressures, including high energy prices, against the risks of disappointing results and a possible collapse in inflated stock valuations.
Concerns were also raised about job displacement and the possibility that countries with limited technological infrastructure could be excluded from the benefits of the AI revolution.
Scientific breakthroughs and new industries
Despite his concerns, Clark presented an optimistic outlook for AI’s role in scientific advancement, arguing that the technology could help researchers tackle complex problems in mathematics, engineering and biology.
He envisaged scientists working alongside hundreds or even thousands of AI assistants to accelerate experiments, develop new technologies and discover medicines.
According to Clark, AI could become a force multiplier for researchers, enabling them to investigate problems that might otherwise require substantially more time and resources.
He also argued that more countries could participate in the AI revolution by investing in the computing infrastructure needed to power advanced systems, particularly data centres.
Such investments, he suggested, could widen access to the technology and create opportunities for countries seeking to benefit from the next phase of digital and scientific development.
Calls for strict oversight of biological research
Clark nevertheless stressed that the same capabilities that could advance scientific discovery might also be exploited for harmful purposes, including the development of biological and cyber weapons.
“Being able to use AI for science opens us up to the misuses of science as well, biological and cyber weapons and things we cannot yet imagine, and we have many choices to make,” he said.
He identified biological research as a particularly sensitive area requiring stringent oversight, noting the potential for advanced AI systems to facilitate dangerous scientific applications.
Biological research, where AI has already created new viruses, is an area “where we want the strictest human oversight and transparency to maintain safety,” he said.
His remarks followed high-profile incidents involving AI systems that reportedly escaped testing environments and gained unauthorised access to websites and databases.
Although Clark maintained that AI remained under human control, he warned that the opportunity to introduce meaningful safeguards was narrowing as the technology became more sophisticated.
He described the remaining period for policymakers to act as “a small number of years”, underscoring the urgency of international efforts to establish rules for the technology.
AI poses challenges for employment
Clark also acknowledged that the growing adoption of AI was already contributing to job losses in some sectors, particularly occupations highly exposed to automation, including call centres and customer service.
However, he said the effects would not be uniform across the economy. In fields such as scientific research, AI could complement human workers and expand their capacity to undertake complex tasks.
A major uncertainty, he added, was whether AI would ultimately generate enough new employment opportunities in emerging industries to offset jobs displaced by automation.
His comments highlighted the difficult choices facing governments as they seek to harness AI for economic growth while addressing its implications for employment, security and social stability.
For Clark, the central challenge is ensuring that the benefits of increasingly powerful AI systems are matched by effective safeguards, international cooperation and human oversight before the window for action closes.