There are few parts of our lives in which AI doesn’t play a part nowadays, whether it is helping us with work tasks, writing letters, or simply helping us to decide where to eat out. It is even commonplace to find people using AI for their own therapy and counselling.
If we are willing to turn to the new technology for such personal matters, perhaps we shouldn’t be surprised that AI-based solutions are increasingly being offered for personal financial planning questions as well. But just because we can, doesn’t mean we should: just how sensible is it to entrust our financial futures to ChatGPT, Claude, Copilot and the rest?
The first thing to consider is that AI is totally dependent on the data it can mine to formulate its ‘thoughts’. And whilst machine learning has certainly improved since the early days, such technology is still not fully capable of distinguishing between quality data and less reliable material.
This can result in ‘hallucinations’ – things which are presented as facts, but which are in fact totally wrong. They are surprisingly common even in the most sophisticated large language models (LLMs). If financial advice is then offered based on these hallucinations, then it is unlikely to be good advice.
Anyone who has used generative AI will know that it is designed never to say ‘I don’t know’, even if insufficient data is available to provide a reliable answer. Instead, it will always provide a response, even if that response is flawed or just plain wrong.
But perhaps the key weakness in this AI-generated advice is that it is not based on any kind of human interaction.
Important though knowledge and experience are in financial planning, the key skill is understanding individual needs and aspirations, and that can only be the result of honest, healthy conversations. Only then can advice about risks, trade-offs and options be offered in a way which genuinely meets the need of each person.
This is because LLMs work by looking at patterns – and these by definition cannot take into account specific individual circumstances.
There is another danger on relying on AI-generated financial advice. When you consult a qualified financial advisor, they are bound by a whole raft of regulation, from protecting your data to ensuring you are offered the best advice. AI offers no such protection; in fact, it works by constantly mining data, and that includes anything you tell it when seeking an answer to financial issues.
As the old adage goes, when you are online, if you are not paying for the product, you are the product – or at least your data is.
All of this is not to say that AI can’t play a role in financial planning. Increasingly it is helping us find information, although there is still a need for a human being to check that information for accuracy.
But whilst financial planning is of course about numbers, it is mostly about people – human beings. And in the end, however good the technology, it will always need a human input.