Lumsdaine's co-authored report is published as an official report of the 16-member Advisory Scientific Committee of the ESRB, of which she is a member. The ESRB is responsible for the macroprudential oversight of the EU financial system and the prevention and mitigation of systemic risk.

• The ESRB flags AI as a potential amplifier of systemic risk in finance
  The report by its Advisory Scientific Committee (ASC) concludes that features common in advanced AI—such as model uniformity, high concentration, over-reliance, rapid execution (“speed”), and barriers to entry—could exacerbate vulnerabilities in the financial system. These AI-driven vulnerabilities may magnify traditional sources of systemic risk (such as liquidity mismatches, common exposures among institutions, lack of substitutability, interconnectedness, and leverage). 
  
  
• AI offers benefits — but those benefits come with trade-offs for stability
  The press release acknowledges that AI tools “have a hugely transformative potential” and may improve tasks like risk management or processing unstructured data. However, the same attributes that make AI powerful in optimization and data handling—speed, model uniformity, widespread adoption—also create new systemic risks if unchecked.
  
• The ESRB calls for regulatory and supervisory adjustments to address AI-driven risks
  To mitigate these emerging stability challenges, the report recommends a “thoughtful and measured policy approach.” Specifically, it suggests revisiting prudential regulation—including capital and liquidity requirements—possibly imposing “skin-in-the-game” or sophistication thresholds for institutions using AI, plus enhanced supervision and stronger competition and consumer-protection policies.

Read the report.