AI use reaches 86 percent of financial firms as Rusu urges safeguards

Rusu said the FCA is immediately embedding six of seven Mills Review recommendations, published in July on retail finance through 2030.

It will build on existing investor tools rather than create a public-interest financial capability tool.

Good and poor practice examples published on October 19 cover consumer outcomes, governance and resilience.

Examples include chatbot vulnerability alerts to human advisers and independent ethics panels; failings include untested journeys and governance added after deployment.

Firms need accountable senior managers, human oversight, stress tests, fallback options and supplier controls.

Rusu said incident reporting must cover major AI failures and near-misses, while agentic systems require continuous, conditional assurance.

Four in five less-experienced investors aged 18-40 use AI for investing, but only 40 percent of consumers recognise the lack of recourse after general-purpose AI advice.

Rusu urged clear provider disclaimers, citing ChatGPT, Gemini and Claude.

InvestSmart, FCA Firm Checker and IOSCO's first TechSprint support investor education.

The FCA is building AI oversight for 60,000 anti-money laundering firms entering its remit in 2028 and about 9,000 small wholesale transaction-reporting firms.

Its Handbook API launched in August; data collection APIs are in development.