TL;DR

Three in four UK pension scheme trustees use AI at least some of the time, according to an Aon survey reported by Employee Benefits on 16 October. Governance has not kept pace. Over half (53%) have had no AI training, 61% are unaware of how their advisers apply the technology, and 94% have nothing about AI written into their contracts.

The survey

Aon polled 61 UK defined benefit (DB) schemes, the smallest with fewer than 1,000 members and the largest above 20,000.

Interest in AI is widespread, but action is thin. Some 65% of respondents had discussed AI opportunities, yet only 18% had recorded any. On oversight, 73% carry out no specific check of how their providers govern AI.

Members may already be ahead of their schemes. While 86% of schemes offer members no AI tools, 94% of respondents believe members are already turning to AI for answers about their pension benefits.

What Aon says

Paul McGlone, a partner at Aon, said the results show schemes “are already working with AI, whether through trustees, advisers or wider member behaviour”, but that use is “relatively ad hoc”. He questioned how clearly trustees can see AI use among the firms that advise and supply their schemes.

McGlone said The Pensions Regulator (TPR) is “due to issue guidance on this area very soon”. He argued the aim should be to make innovation “visible, controlled and accountable” rather than to stop it, starting with mapping where the scheme meets AI, protecting data, challenging providers, and keeping a person accountable for decisions and what members receive.

Looking forward

The figures add a trustee-level view to the Society of Pension Professionals survey we covered in August. In our view, the gap Aon describes is a familiar one: individuals often adopt AI informally before their organisations set rules for it. For trustees, the practical first steps are simple: ask advisers and administrators how they use AI, add AI terms when contracts come up for renewal, and arrange basic training before TPR’s guidance lands.