TPR highlights AI benefits, but issues a note of caution too

The Pensions Regulator (TPR) recently issued their most substantive remarks so far on the topic of Artificial intelligence (AI) in the pensions industry. This captures the rapid shift we have seen across the industry – with the topic now being high on the agenda for most pension schemes and trustee boards. 

TPR’s recent speech

Back in June 2025, TPR stated that trustees “must… understand the role of AI in industry”. 

Now, in a speech on 6 May 2026 (titled “Looking Ahead to 2036: The Future of UK Workplace Pensions”), Nausicaa Delfas (TPR's CEO) addressed the topic in the most detail that TPR has done to this point, stating that:

“AI has already arrived and offers significant opportunities to improve regulatory effectiveness in the future – enabling better use of data to identify risks earlier, target interventions more precisely, and support faster, more informed decision‑making. Used well, it can strengthen supervision, improve outcomes for members, and increase efficiency across the system.

But those benefits come with clear challenges. AI systems must be governed carefully, with strong oversight, transparency and accountability. Data quality, bias, explainability and security are critical risks, particularly in a system built on trust.  AI must support human judgement, not replace it, and its use must always be aligned with our statutory objectives and public expectations.”

Our experience

Since around the start of 2025, we have seen the topic of ‘AI in pensions’ go from a sidepiece at meetings, to now often being centre stage. Trustees are paying increasing attention to AI in particular regarding:

• Capitalising on the exciting benefits it has to offer (such as improved member engagement), and 

• Mitigating the risks (new and existing) that come with it (in particular for members who may rely on misinformation produced by AI tools, and in the form of cyber security). 

TPR’s speech therefore aligns with what we have been seeing. We recommend that trustees explore the benefits of AI, but, as TPR recognises, this should be done “carefully, [and] … with strong oversight, transparency, and accountability”.   

Existing guidance in the area

Along with TPR’s remarks in speeches about AI in pensions, PASA produced some useful guidance on the topic back in October 2025. See here for our summary of that guidance and some suggested actions for trustees and other stakeholders. 

Key takeaway

Whilst the speech does not introduce new rules about AI, it provides insight into TPR’s approach and likely areas of focus. We recommend trustees consider:

• reviewing and updating their governance framework in light of AI (e.g. is an AI policy appropriate for your scheme?), 

• implementing mitigations in respect of risks posed to their members from AI, and 

• engaging with their service providers about third party use of AI.

We work closely with AI specialists from our firm and in the industry to understand how the technology is likely to impact our clients. Please do reach out to Chris Brown, Callum Duckmanton, or your usual Burges Salmon contact if you would like to discuss this further. 

To hear more of what we are seeing in the industry, you can listen to our podcast episode on the topic here. Also, be on the look-out for our guest appearance on the “Pensions and AI Podcast”. 

This article was written by Callum Duckmanton (Solicitor) and Chris Brown (Partner) both in our Pensions & Lifetime Savings Team.

AI systems must be governed carefully, with strong oversight, transparency and accountability... [and it] must support human judgement, not replace it, and its use must always be aligned with our statutory objectives and public expectations. (TPR - 6 May 2026)