Robert Rubin, who served as US Treasury secretary during the late-1990s internet boom, warned that the rapid expansion of artificial intelligence could create financial vulnerabilities, potential job losses and uncertainty over whether massive AI investments will ultimately deliver expected returns.
Speaking at the Greenwich Economic Forum on Tuesday, Rubin said the enormous commitments being made across the AI industry have created increasingly complex financial relationships among technology companies, suppliers and investors.
One area of particular concern is what Rubin called “circularity risk.” Major AI companies are making large commitments to suppliers, while some of those suppliers are borrowing money based on the expectation that the commitments will be honored. That interconnectedness could become problematic if one company is unable to meet its obligations.
“What happens if they can’t fulfill those commitments or all those borrowed against them?” Rubin said.
The possibility of such a breakdown triggering problems elsewhere in the financial system is meaningful, according to Rubin, who previously served as co-chairman of Goldman Sachs. He said the probability of disruption was not “near zero.”
The warning comes during a period of heavy spending on the infrastructure needed to support AI development. Companies developing AI software and constructing data centers have issued large amounts of debt this year. Some market participants believe that borrowing has contributed to rising financing costs around the world.
At the same time, benchmark government bond yields have moved higher, adding another layer of pressure to public finances. The 10-year US Treasury yield reached its highest level since 2002 this week.
Higher yields increase the cost of servicing government debt, intensifying concerns about fiscal sustainability in countries including the United States and France. Rubin said markets appear to be paying greater attention to the US fiscal position than they have in recent years.
“I think what’s happening right now is some realization about our fiscal situation beginning to affect markets in a way it hasn’t for a long time,” he said.
Rubin also pointed to inflation and declining confidence in the government’s ability to address fiscal challenges. Persistent uncertainty surrounding public finances could eventually spill into equity markets by discouraging investment and weighing on productivity, he said.
AI Growth May Come With Labor Market Disruption
The prospect of AI delivering major productivity gains has fueled expectations that the technology could significantly boost US economic growth. Rubin, however, expressed doubt that those gains would be enough to resolve the country’s fiscal problems.
Even if AI leads to stronger GDP growth, he said, the economic benefits could be accompanied by considerable disruption in the labor market.
Knowledge workers may be particularly exposed. Rubin identified lawyers, accountants and television workers among the occupations that could face displacement as increasingly capable AI systems take over tasks traditionally performed by people.
Such changes would create another challenge: helping displaced employees move into new types of work. Rubin said the United States currently lacks effective programs for managing transitions on that scale.
“So could we have higher growth from AI? We could — we probably could well,” Rubin said, while adding that he does not expect those gains to eliminate the country’s fiscal problems.
There is also no certainty that today’s enormous AI investments will generate sufficient returns.
Rubin said some companies could emerge as major beneficiaries of the technology, while others may perform poorly. Whether the overall investment boom ultimately pays off remains uncertain, he said.
Balancing AI Safety With Global Competition
Beyond the financial and economic questions surrounding AI, Rubin also highlighted potential risks from the technology itself.
He described AI safety and climate change as “two massive existential risks,” reflecting concerns about how increasingly powerful AI systems could affect society over the long term.
For policymakers, however, efforts to reduce those risks come with a competitive challenge. The United States must consider safeguards for AI development while also keeping pace with China. Rubin warned that if Beijing does not adopt comparable restrictions, stronger US constraints could put American companies at a competitive disadvantage.
Despite his concerns about the country’s fiscal outlook, political system and technological challenges, Rubin said the United States remains the best place to invest.
Preserving that position, he argued, will require a political system capable of responding effectively to the country’s long-term problems.
“Our political system has to become effective enough, again,” Rubin said. “It doesn’t have to be terrific. It wants to be effective enough to deal with our issues.”