IMF warns AI risks may outweigh benefits without safeguards

Highlights

IMF's Kristalina Georgieva warns that without strong safeguards and ethical frameworks, AI's risks could outweigh benefits and fail to spur quick productivity gains.
Anthropic's Jack Clark highlights emergent EU frameworks and voluntary U.S. commitments for AI standards, but stresses that oversight cannot rely solely on industry self-regulation.
U.S. Treasury's Erin Browne urges the IMF to consider AI's potential positive shocks, while Bank of England's Andrew Bailey signals increased optimism for global alignment on testing standards.

AI governance debate sharpens in Bangkok

As reported by Reuters, Georgieva says during a panel discussion that AI could deliver negative shocks if it fails to generate major productivity gains as quickly as expected. She says the absence of safeguards and an ethical foundation could leave the risks exceeding the opportunities, while arguing that access to digital infrastructure, electricity and regulation is needed for the technology to benefit people broadly.

She compares the pace of AI development to Formula One race cars moving without barriers, saying the central challenge is how to build guardrails for a sector evolving too quickly for traditional oversight. Georgieva backs a mix of industry self-regulation and engagement from major governments, including the U.S., China and India, as well as countries involved in the AI supply chain.

She also says institutions such as the IMF should play a role in areas tied to their mandates, including productivity, labour markets and financial stability. Earlier in the week, the IMF flagged growth and financial stability risks from a possible end to the AI investment boom, alongside high debt levels and a prolonged war in the Middle East.

Officials and industry push for common standards

Anthropic co-founder Jack Clark, who appears on the same panel, says standards for AI are achievable and are already beginning to emerge through early European Union policy frameworks and voluntary commitments in the U.S. He says Anthropic regularly tests its systems, publishes details and shares findings, but adds that oversight should not be left entirely to voluntary action.

Clark says he supports regulation from countries that develop AI and argues that common-sense standards are needed to build confidence in the technology and support economic growth. He adds that Anthropic has brought in third-party evaluators to review data and says more companies should take similar steps.

Other policymakers at the meetings are taking varied positions on the balance between risk and opportunity. U.S. Treasury Undersecretary Erin Browne says in a separate news conference that the IMF is too focused on downside risks and should give more weight to positive shocks such as AI, which the U.S. government is now calling "super intelligence," while Bank of England Governor Andrew Bailey says he is more optimistic than a month ago that countries will align on AI testing standards.

In our earlier coverage of international AI safety coordination, we examined how governments and major developers are weighing stronger safeguards for fast-advancing systems. We noted Nvidia CEO Jensen Huang’s call for cooperation with Chinese AI companies and broader dialogue, including proposals for a U.S.-China incident channel and the UK as a neutral venue for talks on managing cross-border AI risks.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.