LPPA and Burges Salmon survey reveals members still want a human touch when it comes to AI adoption

A survey by LPPA and independent UK law firm Burges Salmon has found that members remain sceptical about using AI to inform decisions affecting their retirement.

The survey, issued by LPPA and conducted in partnership with Burges Salmon, gathered responses from 1,690 members.

It found that while 30% of members are using tools such as ChatGPT, Copilot and Gemini for general use at least once a week, more than half (53%) are not confident in the accuracy of AI-generated information.

For pensions research, AI usage remains relatively low, with only 9% of members using AI tools to research topics such as retirement planning and pensions benefits.

These results suggest members are prepared to use AI as a source of information but not as a replacement for trusted expertise. Only 2% of members said they would be comfortable with AI making pension related decisions without human review, while 61% said they would never be comfortable with this approach.

With human oversight considered, 69% said they were either comfortable now, or could be in the future, with AI supporting pension related decision making as long as a human remained responsible for the final outcome. This included feedback on pension administrators using AI for communications with members.

The major concerns were around lack of human oversight (76%), data protection and privacy (69%), accuracy of information (69%) and trust and transparency (58%), with written feedback highlighting that pensions are simply too important to be left solely to the technology.

The survey also highlighted a knowledge gap around how their personal information was handled, with 39% of members unaware that many AI tools process and store personal data, potentially forever.

Feedback from members, which pension funds and administrators should consider, is that 50% said they either have or would use AI to make a complaint or an Internal Dispute Resolution Process (IDRP). 64% also stated that they would not declare AI usage in the complaint document.

This statistic is of particular relevance, as AI-generated complaints have the potential to increase the volume and complexity of complaints received by pension schemes and funds.

The findings also show members are looking for more guidance, with 76% saying they would find it useful for administrators to explain the risks of using AI when making financial and retirement decisions, and 81% said they would welcome general guidance on using AI responsibly.

Martin Davidson, chief technology officer at LPPA, said: “AI has quickly become part of our everyday lives, and these results shows that while members are interested in the opportunities it offers, they’re clear that confidence, security and access to trusted information remains essential.

“Members recognise the value AI can bring, but they want reassurance that oversight and accountability remain at the heart of important decisions. For LPPA, that’s a positive message. It reinforces the importance of helping members understand new technologies, while continuing to provide trusted pension administration and access to the information they need to make informed decisions about their future.”

Chris Brown, partner at Burges Salmon, said: “With members now using AI on a daily basis and The Pensions Regulator’s guidance on the topic expected soon, now is the time to get up to speed with the opportunities and risks of the technology and to ensure that appropriate governance arrangements are in place.”

Gary Delderfield, partner at Burges Salmon, commented: “This survey shines light on how members are using AI and what their expectations of pension providers are, particularly in the LGPS space. Administering authorities should now be considering what they can do to protect their members, as well as the fund, from the risks created and exacerbated by AI.”

LPPA and Burges Salmon can help pension schemes and LGPS funds to implement the appropriate governance framework on AI. The joint report providing a detailed analysis of the survey results, can be accessed here.