Yellen warns of ‘significant’ AI risks to financial system

Treasury Secretary Janet Yellen warned Thursday that artificial intelligence (AI) could pose “significant risks” to the financial system. 

The “complexity and opacity” of AI models, “inadequate” risk management frameworks, and “interconnections” that arise from using the same data and models can create vulnerabilities, Yellen said at a conference on AI and financial stability. 

Risks can also emerge from concentration among the vendors developing models and providing data and cloud services, as well as from “insufficient or faulty data,” she added. 

“The tremendous opportunities and significant risks associated with the use of AI by financial companies have moved this issue toward the top of Treasury’s and the Financial Stability Oversight Council’s agendas,” Yellen said at the event co-hosted by the council and Brookings Institution. 

As part of her remarks, the Treasury secretary announced plans to launch a formal public request for information about the “current uses, opportunities, and risks of AI in the financial services sector.” 

The Treasury Department’s Federal Insurance Office also plans to hold a roundtable about the use of AI in the insurance sector, she said.

Like Yellen, Securities and Exchange Commission Chair Gary Gensler has similarly warned of the risks that centralized AI could pose to the financial system. 

At a fireside chat earlier this year, Gensler said he thinks there will likely be only a handful of large base models and data aggregators, creating a “monoculture” with hundreds or thousands of financial actors relying on the same data or model. 

“Diversity of models and diversity of data sources; otherwise you end up with a pretty fragile system,” Gensler said at the time.