LONDON, October 9 — The story begins with a proposal from Demis Hassabis, CEO of Google DeepMind, who in a summer 2026 essay suggested the United States create a frontier artificial intelligence standards body modeled on the Financial Industry Regulatory Authority, or FINRA. What unfolded in the essay was a detailed look at how FINRA operates as a self-regulatory organization for securities brokers, overseeing approximately 3,500 brokerage firms without being a government agency.
At its heart, the piece highlighted that FINRA is composed of and funded by the very firms it regulates, with its budget substantially sourced from those member firms and its executives receiving large salaries paid by them. This structure raises concerns about conflicts of interest, as broker-dealer firms elect representatives to FINRA’s Board of Governors, including both industry and public members, many of whom have prior professional ties to broker-dealers.
Financial regulation scholars note the inherent risk in such self-regulatory models: the potential for conflicts when regulated entities grant regulatory powers to an organization representing themselves. Critics describe these models as self-serving, self-interested, lacking in sanctions, beset with free-rider problems, and essentially a sham, with FINRA specifically facing an inherent conflict between its regulatory objectives and the for-profit motives of its member firms.
Some view the securities industry’s self-regulation system as no longer effective in meeting public policy goals, leading other jurisdictions with securities markets to abandon the model due to unworkable conflicts of interest. Enforcement data shows a decline in FINRA’s disciplinary actions: in 2025, it brought the fewest cases in a decade—22% fewer than in 2024—continuing a downward trend since 2015, when it pursued over 1,300 cases, more than three times the 2025 number.
FINRA’s own statistics confirm this decline across multiple metrics, and when enforcement actions occur, fines and restitution imposed on firms are also lower than their peak a decade ago. Possible explanations for the decline include reduced focus on enforcement post-2008 financial crisis, member firms influencing exam and surveillance practices to be less stringent, and a shift in priorities under FINRA’s CEO appointed in 2016, who emphasized engagement with brokers over retail investor protection.
As stated by FINRA’s former head of enforcement, the statistics indicate the organization is more focused on member relations than investor protection. The article cautions that before adopting a FINRA-like model for AI oversight, policymakers should consider whether such a self-regulatory structure can effectively serve public interest given its documented shortcomings in the securities industry. Looking ahead, the debate over how to govern emerging technologies like artificial intelligence will likely continue to draw lessons from financial regulation, with close attention to whether self-governance can truly protect the public good.
