The House of Lords has put the FCA on notice: if ChatGPT can be used to understand insurance, it may be time to regulate ChatGPT and other AI tools.

At a House of Lords Financial Services Regulation Committee hearing in September 2026, a peer revealed AI had helped him make sense of an insurance product. The FCA's Executive Director for Consumers, David Geale, conceded that the boundary between AI guidance and regulated financial advice is becoming "blurry" where AI tools are used for recommendations or specific advice, but confirmed that tools such as ChatGPT, Claude and Gemini remain outside the FCA's perimeter for now.

That position may not hold for much longer. The Mills Review, published on 6 July 2026, names insurance directly as a market AI will reshape, through embedded cover, automated quote comparison and claims triage and guidance.

The FCA has been asked to decide within three to six months whether general-purpose AI models operating outside the regulatory perimeter should stay outside it.

For insurers, the implications are threefold

1. Accuracy and consumer harm

A model may give a customer incorrect information, misread a policy term, or support a decision a firm cannot properly evidence. These are all situations which present a clear risk of customer harm. Around 26% of UK adults already trust tools such as ChatGPT, Claude or Gemini for financial advice, yet only around two in five correctly identify what consumer protection they actually have when using those tools. For insurers, what a model says about an excess, an exclusion, or a claims process is now read by many customers as the answer.

2. Market power and distribution

Control of the AI-mediated customer interface may become a major source of market power, influencing which products are visible, how choices are ranked, and where value is captured. AI is a distribution risk if an insurer becomes invisible in the channels that consumers increasingly rely upon.

3. Regulatory perimeter uncertainty

General-purpose AI assistants sit largely outside the FCA's activity-based perimeter, so a consumer who uses one for a financial decision may have no formal route to redress. If the FCA moves to bring these tools within its remit, even partially, insurers will need to ensure that any AI-mediated distribution of their products meets existing obligations under the Consumer Duty.

Closing comment

This potential direction of travel is interesting. For an industry where a customer's understanding of their own policy has always been a source of regulatory risk, the rise of AI-generated insurance guidance is unlikely to be a problem that can be left to sit outside the perimeter indefinitely.