Former U.S. Treasury Secretary Rubin Warns: AI Boom May Bring Productivity Dividend, But Financial and Social Risks Lurk Behind

Former U.S. Treasury Secretary Robert Rubin warned that the artificial intelligence (AI) investment boom may bring enormous productivity gains, but could also generate financial and social risks that markets have not yet fully priced in. Rubin, who served as U.S. Treasury Secretary during the internet boom of the late 1990s, said he is particularly concerned about "circularity risk" within the AI ecosystem. The term refers to overlapping commitments among suppliers, customers and investors, such as mutual commitments between chipmakers and software companies. "Some of the large AI companies have made enormous commitments, and at the same time, there are a lot of suppliers, and many of those suppliers are borrowing based on those commitments. What happens if those companies can't fulfill those commitments, or if all the borrowing based on those commitments can't be repaid? That's what's called circularity risk," Rubin said Tuesday in an interview at the Greenwich Economic Forum. Rubin said some disruption risks are realistic for example, one party failing to honor its commitments and triggering a chain of defaults. He believes such risk is not "close to zero." The former Goldman Sachs co-chairman said this scenario is indeed possible. This year, companies building data centers and developing AI software have issued debt on such a massive scale that many market participants believe it has driven up global borrowing costs. This week, benchmark government bond yields climbed further, with the U.S. 10-year Treasury yield touching its highest level since 2002. This in turn has pushed up government debt-servicing costs, intensifying concerns about the fiscal sustainability of the United States, France and other countries. When Rubin joined the Bill Clinton administration, markets were already broadly worried about budget deficits. He said markets are now beginning to show anxiety about large-scale borrowing. "I think what's happening now is that the market is beginning to recognize our fiscal situation, and that recognition is starting to affect the market in a way that hasn't been seen for a very long time," he said. He also pointed to inflation and an overall decline in market confidence in the government's ability to address fiscal problems. "When you face an unfavorable or unstable fiscal situation, it can further affect market confidence in the government's ability to deal with it," he said, adding that this dynamic could ultimately affect the stock market as well, because uncertainty tends to dampen investment and productivity growth. Productivity Doubts Rubin pushed back on the view that AI-driven productivity growth can allow the United States to effectively grow its way out of its fiscal troubles. The former Treasury secretary said that while a productivity acceleration could indeed boost GDP growth, those gains could come with significant job losses, especially among white-collar workers. He predicted that "this will have a very severe impact on knowledge workers." He said lawyers, accountants and people working in the television industry could all face the risk of being replaced, which would raise a difficult question: how should people transition into new jobs? He noted that the United States currently lacks an effective plan to deal with this kind of employment shock. "Can AI bring higher economic growth? Yes it very likely can. But I don't think it will solve the fiscal problems we have to solve," he said. He also questioned whether massive investment in AI will ultimately generate sufficient returns. "Will these investments ultimately pay off, or not? I don't know, but neither does anyone else," he said. "Some of these investments will end up producing very good results, some companies will do very well, and a large number of others will do poorly." AI Safety Concerns Rubin said his concerns are not limited to markets and the economy, but extend to the security risks posed by increasingly powerful AI systems. He said he considers AI safety and climate change to be "two enormous existential risks," different in nature from any risks humanity has faced before. Rubin also noted that the United States faces a dual challenge: trying to put safety guardrails around AI while also having to focus on competition with China. Rubin also said, "The United States is still the best place to invest." But he believes maintaining that advantage depends on whether the U.S. political system can become effective enough to address fiscal, technological and other long-term challenges. "Our political system has to become effective enough again and right now it can't do that. It doesn't need to be outstanding, but it has to be effective enough to deal with the problems we face," Rubin said.