Should UK financial regulators take a more proactive stance on AI risks?

There was no real fracture here, only gradations of conviction, and the vote settled the question of whether regulators should lean in on AI risk by a margin of roughly four to one. What's striking is that the strongest consensus came not from the digitally fluent young but from the oldest voters in the chamber, who backed a more proactive stance almost to a person, while the youngest cohort, presumably closer to the technology day to day, showed relatively more of the dissent that did exist.

That pattern cuts against the easy assumption that unease about AI is generational nostalgia for a simpler world; instead it reads more like accumulated caution about financial instability from those who remember what happens when oversight lags the market. A gender gap ran through the numbers too, with women somewhat more hesitant than men, though far from opposed, suggesting the debate splits less along who fears AI and more along who fears under-regulated finance most.

The result lands squarely on the fault line the Treasury Committee itself identified this year: whether the "stabiliser not a brake" model still holds when the risk in question moves at machine speed, or whether the Financial Services AI Adoption Plan's growth ambitions need firmer guardrails first. It echoes a familiar British instinct on regulation generally, that caution is best applied before the crisis, not audited after it.

The floor's top Yes case put it starkly: AI "can go horribly wrong astonishingly quickly," making guardrails "an essential safety measure." The most-respected No voice countered with a harder-edged institutional scepticism, noting regulators "wouldn't trust them to properly regulate the last crisis" and were "unlikely to get the next one right" — a dissent less about AI than about faith in regulators at all.

If there was a consensus here, it was less enthusiasm for AI regulation than a lingering distrust of what happens when nobody minds the machine.