Do 28 million people in the UK really use AI for personal finance?

According to a recent report by Lloyds Banking Group "More than 28 million UK adults are now turning to artificial intelligence to help manage their money – making personal finance the nation’s number one use of AI."

There are two big findings there, and I was intrigued to find out how Lloyds got to them.

As part of their 2025 Consumer Digital Index, Lloyds conducted telephone, face-to-face and online surveys.

Out of 5,000 "digitally engaged individuals" who responded to the online survey, 56.2% said they had used AI tools for personal finance and money management within the last 12 months. This was much higher than the next highest category: 28.6% said they had used AI tools within the last 12 months for writing emails, letters or work documents.

Lloyds say that extrapolations in the report "use the latest available 18+ UK population estimates released by the Office for National Statistics."

So what does it mean?

The "more than 28 million" appears to have been arrived at by applying the 56.2% ratio from the survey to the UK adult population. The 5,000 respondents were "representative (based on age, gender and region)". But given that they are described as "digitally engaged", are they representative of the UK population in AI use?

The survey data also shows that 42.8% of the respondents were "very engaged" with their finances. Are they drawn perhaps from a section of the population that are more inclined to use AI for their personal finances?

While the survey responses show that personal finance is the highest use case for AI tools within the 5,000 respondents, a straight extrapolation to the whole UK population feels like something that could be open to challenge.

But let's assume for a minute that 28 million people are using AI for personal finance. Based on the survey we'll also assume that the majority are using generative AI (the study said 60.3% of those using AI had used ChatGPT). Is this something that it's well suited for?

Remember what AI really is

Generative AI like ChatGPT doesn’t understand money. It predicts what words are likely to come next. When you ask it a question about savings, mortgages or investments, it isn’t considering your goals or assessing risk. It’s drawing on patterns in language that have appeared in similar questions before.

That’s why its answers sound confident. Fluency is what these systems are built for. Understanding is not. AI can write fluently about money and be completely wrong. Like the AI-written articles a publisher had to remove because they made fundamental mistakes in explaining compound interest.

Context is always essential when using AI, and becomes critical when discussing money. Financial choices need to be made in relation to your age, income, debts, goals, appetite for risk. Errors will cost you money, and the AI model has no oversight or liability like a regulated adviser.

Financial institutions' own AI products and specialist financial AI may have some more context about your situation and be subject to a level of oversight, but fundamentally they are still models without a true understanding of the impact their recommendations can have for you.

Set the right intent, question everything, and use your judgement

If you choose to use AI to research options and explore possibilities, it's essential that you set out your intent as fully as possible. Be clear about your circumstances, what matters to you, and what your goals are. Explain your appetite for risk.

Question everything in the response rather than assuming that it is factually correct. Use other sources to confirm where necessary. Consider the assumptions that have been made, and use your common sense to check them.

Finally, use your human judgement to assess whether any recommendation is right for you. If the recommendation is wrong, you are the one who will lose out. So make sure that you've arrived at a decision through your own thinking, with the AI supporting your research rather than deciding for you.