The Financial Conduct Authority is at risk of missing an opportunity to prevent “widespread consumer harm” from unregulated artificial intelligence tools, the Investment and Saving Alliance has warned as a consultation on simplified advice closes today. The organisation wants the FCA to introduce guardrails into unregulated AI tools and police the boundary between regulated and unregulated advice in AI – or relax requirements on regulated advisers so they can compete better. 

The consultation which closes on Friday highlights two-year-old government findings that 73% of the public have incorporated AI into their daily lives within the last month, but only notes that this means “advice firms will need to adapt to changing consumer expectations for how they receive financial advice”.  

When it comes to AI, the FCA currently seems to focus its attention on cyber security rather than unregulated advice, with reports of new models like MythosAI being capable of quickly finding and exploiting vulnerabilities in legacy systems. The regulator recently made a statement on frontier AI models and cyber resilience jointly with the Bank of England and the Treasury.  

However, the Mills Review, which the FCA launched in January, includes exploring the impact of AI on consumers, such as how consumers will be influenced by AI and how financial regulators may need to evolve to ensure that retail financial markets work well. Sheldon Mills will report with recommendations in the summer.

TISA said the FCA’s proposals on advice and guidance in the simplified advice consultation fail to grapple with the realities posed by AI, urging the watchdog to safeguard consumers and industry from unregulated AI tools conducting unauthorised financial activity without any guardrails. 

While AI advice-like services democratise financial information and support, consumers might end up acting on “unsuitable recommendations with no clear accountability, suitability assessment or route to redress”, warned Sophie Legrand-Green, head of policy for consumer protection and access.  

“We urge the FCA to work with AI providers to build appropriate guardrails and signposts into [large language models], warning users about their limitations and the lack of regulatory protection and directing consumers to reliable, regulated forms of support and advice,” she said.  

Legrand-Green added that unregulated, free AI tools risk undermining regulated advice and targeted support by “offering personalised-sounding answers at scale, while regulated firms remain constrained by rules that make new advice models commercially difficult”. 

She called for the advice-guidance boundary to be policed consistently – including for AI tools – saying if that is not possible, regulated firms will need a framework that allows them to compete safely and serve consumers at scale. She blamed charging and qualification requirements for regulated advisers being unable to compete with AI.