Anthropic co-founder Jack Clark, speaking at the IMF and World Bank annual meetings in Bangkok, said rapidly improving AI “will bring far faster change across all fields,” with one of Anthropic’s internal economic models forecasting that AI could expand US economic output to about $44 trillion by 2030, Reuters reported (David Lawder; English text via CNA’s Reuters syndication, datelined Bangkok, 17 October 2026).
That $44 trillion figure is not an IMF forecast and not a Reuters projection. It is what Clark described as an Anthropic internal model output. Reuters notes it compares with a current US level of just over $30 trillion. Clark himself said the estimate “sounds very outlandish, I know.”
What Clark said the forecast assumes
Per Reuters’ account, Clark tied the number to “the technology continuing to get better,” adding that if GDP grew that much, “economic doctrine suggests that new industries and inventions would need to have been created to power such a major economic expansion.”
He also called for international cooperation to erect “unusual and serious rules” for transformative AI, and said the window for policymakers to act is shrinking. Reuters quotes him telling hundreds of government and finance officials that the pace of progress in AI and science “will speed up, and this will feel confusing and disorienting, and will bring with it great wealth and potentially great peril.”
On misuse, he warned that using AI for science also opens “misuses of science as well, biological and cyber weapons and things we cannot yet imagine.” Biological research, he said, is an area “where we want the strictest human oversight and transparency to maintain safety.” He said AI is still under human control for now, but the window to impose safeguards is “a small number of years.”
Jobs — briefly, and separately from other Anthropic papers
Clark acknowledged job losses in some “high AI exposure” fields such as call centres and customer service, while arguing that in scientific research AI can enhance employment, with a major unknown being jobs in entirely new industries. That is Clark’s public remarks, not the internal employment telemetry in Anthropic’s separate “How employment is changing at Anthropic” paper.
Do not blend with the expectations-gap growth rates
TSN’s expectations-gap survey post covers staff and public views, including model-implied GDP growth rates of 5.4% and 17.1% from that survey exercise. Those percentages are not the same object as Clark’s $44tn-by-2030 internal-model sketch. Keep them in separate stories; do not average or rewrite one into the other.
For related White House process context, see TSN on the Super Intelligence Force charter.
What’s still unknown
- The full specification, assumptions and confidence bands of the internal model Clark cited
- Whether Anthropic will publish that model’s methodology
- How finance ministers in the room treated the $44tn line — Reuters reports the speech, not a formal IMF endorsement
What this means
Clark used an Anthropic in-house model to put a vivid ceiling — about $44tn US GDP by 2030 versus just over $30tn now — in front of IMF/World Bank audiences, while calling the number outlandish and urging “unusual and serious rules.” Report the figure as his company’s model, as he described it, not as an official international forecast.
Sources
- [1] CNA / Reuters (David Lawder): AI rapidly improving, could vastly boost US GDP, Anthropic’s Clark tells policymakers, 17 October 2026
- [2] Reuters syndication via U.S. News (same report)
Related stories
- Anthropic Survey Finds a Wide AI Expectations Gap Between Staff and US Adults
- White House Super Intelligence Force: what the charter says
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
