The Financial Conduct Authority (FCA) has issued a warning to young investors following research findings which showed four in five 18- to 40-year-olds have used AI to inform their investment decisions. Around two-thirds said they utilise the technology in this way either occasionally or regularly.
Over half (56%) said they trust AI tools more than TV and radio (47%), press (46%) or social media influencers (29%).
In addition, despite 73% acknowledging that AI can provide inaccurate information, 38% believe it’s fine to make an investment decision based solely on AI outputs.
Separate research by Quilter found that 42% of a cohort of 2,002 UK-based adults feel comfortable using AI to explain financial concepts, 40% using it to help understand tax rules and 29% to sense-check their thinking.
Young investors are also turning to AI to advise on their pensions.
Standard Life research found that 39% of surveyed 18- to 34-year-olds have already used AI to get information about pensions or saving for retirement, compared with 16% of 35- to 54-year-olds and just 6% of those aged 55 or over.
Trust in AI is misplaced
The watchdog warned that such wholesale trust in AI is misplaced, explaining that common beliefs that AI-generated financial information is regulated and that the Financial Services Compensation Scheme (FSCS) or Financial Ombudsman Service (FOS) will provide compensation if AI advice goes wrong are simply not true.
Lucy Castledine, director of consumer investments at the FCA, said: “AI can help you research companies, understand jargon or explore options before you make a decision – but you need to understand how you are protected and continue to use your own judgement.”
The FCA has urged more inexperienced investors to double-check the sources utilised by AI tools, understand that AI cannot predict future investment trends or performance and be clear that there is no safety net for AI-generated advice.
Dan Coatsworth, head of markets at AJ Bell, said: “The FCA research highlights that many younger people are putting faith in AI for investing.
“It would be devastating if they made major investing mistakes by trusting AI as that could dampen their enthusiasm for saving in the future as people having a bad experience might become less willing to put away as much as possible and that could lead to a poor quality of life down the line.”
A human professional, such as a financial adviser, is also better equipped to adjust their advice according to the person sat in front of them, considering their age, gender, life goals and risk appetite.
“There is a risk that AI systems are making assumptions based on broad circumstances,” Coatsworth said, noting that “one person’s attitude to risk and their investment timeframe might be completely different to the next person”.
AI is also still learning, collating information from a wide range of sources – the credible and not so credible, therefore providing information that may not be correct, he added.
The Quilter research further highlighted an additional risk: the sharing of personal information. More than one in five (22%) of survey respondents said they use AI to get more accurate recommendations via personal information, with comfort doing so greatest among the under 45s (37%).
Three ways to navigate AI and your personal finances safely
The reality is that, as AI gets more sophisticated, it will become a more ingrained facet of peoples’ lives. This includes when making financial decisions.
As such, Sam Christopher, proposition direct at Quilter, has provided three tips to navigate the use of AI in financial and investment planning.
Context matters
Put simply: the output from AI is usually only as good as the prompt itself. Christoper noted that the provision of important context will better ensure AI doesn’t provide vague or irrelevant information.
“Especially when asking about tax or portfolio construction, being specific about things, even if just hypothetically, and what you are trying to achieve will likely result in a better outcome,” she said.
She provided the following prompt examples:
Good practice: "Imagine you're helping a 49-year-old woman in the UK who wants to retire at 67. She has a workplace pension and would like to know whether increasing her monthly contributions could improve her retirement income."
Bad practice: "How much should I save for retirement?"
Personal information must remain under wraps
Context is important but this does not mean sharing personal information that should stay private, such as bank account details, passwords, home addresses or security information.
“Many good prompts can be constructed without this information, and as the technology advances at pace, privacy concerns will continue to exist,” Christopher said.
Good practice: "I'm looking for general information about pension tax relief for someone in their 40s earning around £50,000 a year."
Bad practice: "My name is John Smith, I live at 1 High Street, my date of birth is 1 January 1977 and my pension account number is 12345678. How much tax relief can I claim?"
Always double-check
What an AI tool says should never be treated as gospel.
Christopher said: “AI is well known for its hallucinations and can sound confident and authoritative, even when it’s wrong.”
As such, she said it should be seen more as a starting point for research than a source of definitive answers.
“Don’t be afraid to challenge the technology and interrogate the sources it has used, then compare this across multiple sources and check it rings true,” she said.
Good practice: "Can you explain the pension annual allowance and tell me which sources you've used?" Then check the information against a trusted source such as GOV.UK or HMRC.
Bad practice: “Can you explain the pension annual allowance?” Then take action with your pension, using the information as fact.