Ask a chatbot whether you should move your pension and it will tell you. It will tell you nicely and that it’s a great idea – it wants to please you!
It won't ask if you can afford to be wrong or give you anyone to call if you are. Under British law today, it doesn't have to.
That gap, between what the machines already do and what they answer for, is the real story here.
Governments, on the whole, have chosen not to referee this. The UK’s 2023 AI white paper set out a light-touch, principles-based approach rather than a dedicated rulebook, so a general-purpose AI tool dispensing views on your pension sits entirely outside the financial services realm.
The firms that have always given financial advice don’t have that luxury. They remain inside it, fully. The same question and the same worried saver meet two entirely different levels of protection depending on who – or what – answers.
The market already prices this in
UBS thinks the disruption is real enough to put a number on it. The bank has built AI-driven competition into its base case for St James's Place, assuming gross inflows plateau and net flows eventually turn negative, with stress scenarios implying downside of 20 to 60%.
In February, SJP's shares fell by roughly 12% in a single session on precisely this fear. Markets already believe AI will take a bite out of advice businesses. What should trouble a saver is that only one side of that contest carries any accountability for getting the answer wrong.
What accountability buys you
Being regulated means five concrete things rather than an abstract sense of trustworthiness.
1) A firm has a legal duty to act to deliver good outcomes for you, under the Consumer Duty.
2) It has a named senior individual personally accountable for that outcome, under the Senior Managers & Certification Regime. They can go to jail if they get it wrong!
3) Every promotion it sends you has been checked and signed off before you saw it, under section 21 of the Financial Services and Markets Act 2000.
4) If something goes wrong, you can complain to the Financial Ombudsman.
5) And if the firm itself can't pay redress, the Financial Services Compensation Scheme steps in behind it.
An AI tool answering the same question has none of this. Not one of the five.
A very real trade-off
People like to describe regulation as a tax on incumbents – the thing that lets nimbler entrants move faster. It is a tax, I won't pretend otherwise.
Every promotion we send is approved before it goes out. Every outcome has a named individual answerable for it under SM&CR.
Meanwhile, a chatbot with none of that structure can produce an answer in the time it takes you to type the question. That's a genuine disadvantage – and it isn't going away.
I'd rather compete on speed without giving up the accountability that comes with it: build fast inside rules that exist for good reason and say plainly what people are getting when a machine answers a question about their money.
Something will eventually go wrong. It always does. The question that matters is who picks up the pieces.
Ask a chatbot and the answer is you – on the hook 100%, whether or not you realise it going in. Ask a regulated firm and the answer is us.
That’s what the fee difference is actually paying for: not a nicer interface than Anthropic’s or OpenAI’s, but someone answerable when it counts.
AI can make good investing cheaper and easier to reach. It’s already doing that, and on balance, that’s a good thing. What it shouldn’t become is a route into financial advice from something that owes you nothing when it gets the answer wrong.
This article is for informational purposes only and does not constitute personal financial advice. If you are unsure whether an investment is right for you, please seek regulated financial advice.
Sources: Financial Times, 7 July 2026; Proactive Investors, 3 July 2026; Citywire New Model Adviser, 10 February 2026; FCA Consumer Duty (PS22/9); Senior Managers & Certification Regime; Financial Services and Markets Act 2000, s.21.