AI models give inaccurate financial advice 57% of the time, Saturn finds
Answers included calculation errors, missing risk warnings, overlooked tax changes and references to rules that did not exist.
Research from Saturn found that widely used artificial intelligence (AI) models gave inaccurate answers to financial questions 57% of the time.
The report, Artificial Authority: Should you trust AI to deliver financial advice?, tested 18 models from providers including ChatGPT, Claude, Copilot, Grok and Gemini.
The models gave accurate answers in 43% of cases.
The error rate rose to an average of 88% for harder questions, with some models answering 99% of complex questions incorrectly.
Answers included calculation errors, missing risk warnings, overlooked tax changes and references to rules that did not exist.
Saturn said one error involving pension tax rules could have left a saver facing a £17,500 charge from HM Revenue & Customs.
Free models made mistakes in 63% of answers, compared with 49% for paid models.
On the hardest questions, free models returned incorrect answers in 93% of cases.
Researchers also identified errors relating to debt, student loans, mortgages, pensions, tax and savings.
Some models recommended paying the highest-interest debt before priority bills such as rent and council tax.
Saturn said following this advice could put people at risk of eviction, bailiff action or legal proceedings.
Claude also stated that a graduate moving abroad could stop student loan repayments, while a Gemini model said a mortgage payment holiday would not affect a borrower’s credit score.
The report followed research from the Financial Conduct Authority (FCA) which found that 26% of consumers trusted general-purpose AI tools for financial advice.
Amal Jolly, CEO at Saturn, said: “The low quality of financial advice from mainstream AI models risks leading to widespread consumer harm.
“Millions of people are trusting the AI models for money advice, but they are getting wrong answers that can lose them money.
“AI tools are fast becoming the first port-of-call for many consumers with financial questions.”
Jolly added: “But our research reveals evidence that it is far too early to put so much trust in AI chatbots.
“We found some of the most widely used models gave the wrong answer to financial questions the majority of the time.
“In the worst examples, relying on AI’s answers to tax questions could cost families tens or hundreds of thousands of pounds.”
He said: “AI financial advice is currently unregulated, leaving consumers with none of the protections, including compensation, that they would get if they went to a human adviser.
“The FCA has started to think about this, but it needs to act fast to protect people.
“The FCA should regulate AI to ensure consumers are protected.”
Related
Financial AI must consider wider customer circumstances, expert warns
13% of people use AI tools for personal finance advice – Capital One UK
Mortgage Lab launches AI training platform for mortgage advisers
Rightmove adds AI-powered Comparison Assistant to search results
Reapit named 2026 WIRED Trailblazer for innovation in estate agency technology
SPF Private Clients adopts LendWell AI mortgage platform
88% of advice firms put client outcomes at centre of technology strategy, study finds
Aviva expands AI underwriting tool across protection suite
Curvestone AI partners with Connect for Intermediaries to automate compliance checks
Nivo launches AI case coordination tool for specialist lenders
Scottish Building Society goes live on SBS Sovereign AI platform
