UK regulator may bring ChatGPT, Claude under its rules
According to the review, more than one-quarter of UK consumers believe that general-purpose models like OpenAI's ChatGPT
Britain's Financial Conduct Authority should examine whether general-purpose AI tools like ChatGPT, Claude, and Gemini need to fall under financial regulation, according to FCA Executive Director Sheldon Mills, as consumers increasingly turn to them for money decisions.
Mills made the recommendation as part of the newly published Mills Review, a long-term look at how AI could reshape retail financial services through 2030 and beyond.
According to the review, more than one-quarter of UK consumers believe that general-purpose models like OpenAI's ChatGPT, Anthropic's Claude, and Google's Gemini provide financial advice, with very little recognition that the safeguards protecting customers in the area of financial services regulation do not apply to these models.
Mills recommends that the FCA make a determination in the coming three to six months about whether it should expand its perimeter to address the nature, scale, and impact of these general-purpose models, which at present fall outside of its jurisdictional reach.
Chairman of the FCA, Ashley Alder, noted that the FCA's current principle-based regulation approach must be able to match the speed of the developing technology.
The findings of Mills' review reflect this concern in a wider context in that the fast-moving AI deployment process, from cyber and operational risks associated with frontier models through to challenges posed by agentic models, is putting pressure on regulators across the globe, not just the UK.
Aside from consumer-level risks, a different issue came up through this review. The dependency of financial firms on few technology vendors for vital operations was highlighted.
According to a recent global survey, 81% of financial firms are using AI in some form, while 40% have scaled to an advanced level in their use of AI.
Current uses of AI still rely mostly on back-office functions that have low levels of risk, yet financial firms in Britain are increasingly using AI to interact with consumers, even in managing customers' complaints and giving investment advice.
This warning from Mills' review is that dependence on the same models, the same cloud vendors, and similar infrastructure can lead to correlation, herding, and points of failure in the financial system.
-
Anthropic CEO Dario Amodei breaks silence on open-weight AI ban debate
-
Meta accused of disregarding teen harm research in landmark child safety trial
-
China reaches major milestone as domestic DUV chip tools enter mass production
-
China vs US tech rivalry escalates as Beijing accuses Washington of ‘AI hegemonism and distillation’
-
Snap Map rolls out Now Playing feature: Check details
-
OpenAI to expand Dublin workforce threefold as European operations grow
-
Former Amazon scientist bets on cheaper AI to disrupt voice race
-
Nvidia forms industry-wide AI security alliance after Hugging Face hack