Curious but cautious: AI and pension trustees
I recently joined a session on AI in pensions with contributions from The Pensions Regulator (TPR) and Future Focused Advisory. The message was “curious but cautious”: AI offers real opportunities, particularly in administration, but only if adoption is safe, purposeful and proportionate.
Why this matters now
An April 2026 SPP survey found that every respondent from the industry was using AI, up from 87% in 2025. Savers are using it too, including when preparing complaints under IDRP. Even if your scheme is not an advanced user, you will encounter AI.
What TPR expects
TPR is cautiously optimistic. It wants trustees to start with the problem to be solved, treat AI as a tool rather than a decision-maker, and put governance in place before or alongside implementation. Trustees remain accountable even where an administrator or adviser is the one using AI. Principles-based guidance for trustees is due during 2026.
Six questions to ask
The session also introduced a proposed industry guide, which is not TPR guidance and is not endorsed by TPR. At its core are six questions for administrators and advisers:
-
Where does AI touch our scheme?
-
Who checks it, and is the check meaningful?
-
Where does member data go?
-
What happens when something goes wrong?
-
Does any AI act autonomously?
-
Does the provider’s insurance cover AI-related errors?
Ask for written, evidenced answers, and apply the level of scrutiny proportionately to your scheme and the use in question.
My takeaway
Trustees do not need to become AI experts. They do need to ask informed questions, challenge vague assurances and keep human responsibility and member protection at the centre. Standing still carries its own risk, but adoption should never run ahead of governance.
Reach out to sean.burnard@lawdeb.com to find out more.
Loading carousel...