FSB, FINRA and Canada's OSFI have each flagged AI as a threat to financial stability, citing stretched valuations, rising retail leverage and cyber vulnerabilities
Some of the world's most influential financial regulators are sounding the alarm on artificial intelligence; not just as a compliance challenge for individual firms, but as a structural threat to the stability of global financial markets.
In a letter to G20 finance ministers and central bank governors on August 28, Andrew Bailey, chair of the Financial Stability Board and governor of the Bank of England, identified frontier AI as one of the most pressing emerging risks to the international financial system.
Read: As AI upends financial services, could governance become an advantage?
The warning arrives as AI-related asset valuations are described by the FSB as "elevated" and as retail investors are pouring into leveraged products tied to the sector.
"Frontier AI may have the ability materially to alter the speed, scale and economics of cyber risk, which could undermine market confidence system-wide," Bailey wrote in the letter.
A convergence of risks
In Canada, the Office of the Superintendent of Financial Institutions and the Financial Consumer Agency of Canada have jointly documented the scale of the risk.
Their report, which draws on data showing AI adoption among financial institutions rising from roughly 30% in 2019 to 50% in 2023, with 70% adoption projected by 2026, warns that the sector cannot treat AI risks as optional to manage, even for institutions not yet deploying the technology.
The OSFI-FCAC report notes that the International Monetary Fund estimates 60% of jobs in advanced economies will be affected by AI automation, and that severe cyber attack costs on financial institutions have quadrupled in recent years.
Ninety-one percent of financial institutions are reconsidering voice-verification systems in response to AI-enabled voice cloning, according to a Biocatch survey cited in the report.
A July 2024 global IT outage used in the report as a case study in third-party concentration risk, resulted in approximately $5.4 billion in financial losses, illustrating the systemic exposure that can follow from dependence on a small number of shared technology providers.
The FSB's concerns are threefold. First, it flags the financial stability implications of concentrated AI-sector investment, noting that cross-investment between AI companies and large cloud and computing providers (so-called hyperscalers) creates amplification risk if valuations correct.
Second, it identifies a rise in leveraged trading strategies, including among retail investors, as a feature of what it calls "a maturing financial cycle."
Third, and most urgently, it warns that many jurisdictions lack the regulatory protocols to manage the development and deployment of advanced frontier AI models.
"Rising leverage is a feature of a maturing financial cycle," Bailey wrote. "While it can reinforce rising markets, it can also intensify declines."
Advisors are already fielding questions about pre-IPO AI exposure and the risks of buying into high-profile listings at inflated valuations. The FSB's letter adds a macro-prudential dimension to those conversations, that the risk is not just at the portfolio level, but potentially systemic.
FINRA puts US broker-dealers on notice
In its 2026 Annual Regulatory Oversight Report, the Financial Industry Regulatory Authority made clear that the proliferation of generative AI tools across broker-dealer operations raises firms' compliance obligations.
FINRA flagged concerns about data quality, model bias, and the governance of AI-generated client communications, requiring that firms treat AI-enabled interactions with the same supervisory rigor as any other client-facing activity.
The report, published in January 2026, explicitly warns against reliance on limited or outdated datasets, which can produce skewed outputs in client-facing tools, a particular concern in an environment where AI is rapidly reshaping how advisors work with clients and how they communicate investment recommendations.
The FSB has called on global authorities to establish coordinated protocols for the safe release and deployment of frontier AI models, and to strengthen recovery capabilities in the event of simultaneous disruption across multiple financial institutions.
Whether those frameworks materialize quickly enough to match the pace of AI development is a question regulators across multiple continents are now asking aloud.
