The Financial Conduct Authority is seeking to bring regulation to life by providing examples of good practice in the adoption of artificial intelligence as well as details of areas in which regulated businesses could do better.

It has published a guide with case studies detailing best practice and examples of where firms will need to reappraise their adoption of AI to ensure its use does not compromise customer outcomes.

The FCA says firms that use AI safely and responsibly are those that define the consumer need and target market carefully before deployment. They test consumer outcomes and use proportionate controls to manage AI bias and accuracy. They tell consumers how they have used AI and monitor live outcomes.

Areas for improvement include AI tools built without establishing the consumer problem, target market or foreseeable harms. The regulator also warns against guardrails that address one risk but create another and projects that exclude consumers in vulnerable circumstances by default.
Cormac Bradley, senior actuarial director at financial services consultancy Broadstone, said: “AI deployment is impacting many functions within insurers’ business models, across the full spectrum of front and back office, each of which introduces different types of risks. AI governance will have to distinguish between these use cases and introduce appropriate controls and risk management for each.

“The FCA’s paper on good practice and areas for improvement highlights the importance of these where AI can impact on Consumer Duty, particularly where it is being used to inform decisions that could have significant financial consequences for customers.”