IMF chief Kristalina Georgieva said on Friday that the risks of artificial intelligence could outweigh its benefits if no adequate safeguards and ethical framework are established. Speaking at the annual meetings of the IMF and World Bank in Bangkok, she backed a mix of self-regulation by the industry and input from key governments.
AI was a leading topic at this week’s gathering. On Tuesday, the IMF had warned of significant risks to growth and financial stability, pointing to a possible end to the AI investment boom, high debt levels and a prolonged war in the Middle East.
Calls for guardrails and an ethical foundation
Speaking on a panel that included Anthropic co-founder Jack Clark, Georgieva said AI was already strengthening the global economy. She warned, however, that it could also cause negative shocks if it failed to produce major productivity gains as quickly as people hoped.
“If there are no safeguards, then the risks may exceed the opportunities. And above all, we do need an ethical foundation,” she said, adding that she was not sure society had reached that point yet.
On a separate panel, Bundesbank President Joachim Nagel said AI could widen global imbalances, pointing to the large investment needed for both hardware and software development. Georgieva said several preconditions must be met before people everywhere can benefit from the technology, including access to digital infrastructure, electricity and adequate regulation.
She said the central challenge now was how to build guardrails for a technology evolving so quickly. She compared AI developments to Formula One race cars speeding around without barriers to protect spectators.
Debate over self-regulation and government oversight
Georgieva said industry self-regulation was appropriate, since governments were not well placed to regulate or oversee a sector that changes constantly. She added that such self-regulation should involve engagement with society. The countries with the largest AI sectors, including the United States, China and India, along with countries in the AI supply chain, should take part, she told the panel.
Clark picked up the racing comparison, joking that it was “Formula One at night, without maybe brakes in the cars.” He argued that setting standards was possible and said early efforts were already under way, including policy frameworks in the European Union and voluntary commitments in the United States.
Clark said his company regularly tested its systems, published details of those tests and shared the information. He did not believe the process should be entirely voluntary. Asked whether he would favour regulation by the countries that produce AI, he said: “Yes. We need common sense standards, and that’s how you ultimately give confidence in the technology and build the economy.”
Competing views on AI’s economic impact
Harvard economist Dani Rodrik said it was unclear whether the productivity benefits of AI would materialise, calling the situation a double-edged sword. If the gains arrive, they could bring substantial job losses and displacement in affected sectors. If they do not, heavy spending on data centres could set off a costly financial correction, he said.
US Treasury Undersecretary Erin Browne said the IMF focused too heavily on risks to the global economy and should give more weight to potential positive shocks, such as AI. She noted that the US government now calls the technology “super intelligence.” On Thursday, she said Washington was discussing the new name with other IMF shareholders for a joint statement being drafted by the fund’s steering committee.
German Finance Minister Lars Klingbeil took a different view, saying: “I will continue to call it artificial intelligence.” Bank of England Governor Andrew Bailey told a meeting of the Institute of International Finance that he was more optimistic than a month earlier that countries would agree on testing standards for AI. He joked about the naming debate, referring to “the artist formerly known as AI.”
Federal Reserve Chair Kevin Warsh also joined in, telling Georgieva during a separate panel: “Kristalina, you pronounce it AI. In the US we now pronounce it SI.”
