Pension scheme members remain sceptical about using AI to inform decisions affecting their retirement, a survey has found.
The research, by Local Pensions Partnership Administration (LPPA) and law firm Burges Salmon, gathered responses from 1,690 members.
While 30% use tools such as ChatGPT, Copilot and Gemini at least weekly, 53% are not confident in the accuracy of AI-generated information. Only 9% use AI to research pensions topics.
Just 2% would be comfortable with AI making pension decisions without human review, and 61% would never be. However, 69% were comfortable now, or could be in future, with AI supporting decisions provided a human remained responsible for the outcome.
The biggest concerns were lack of human oversight (76%), data protection (69%) and accuracy (69%). Some 39% were unaware that many AI tools process and store personal data.
The survey also suggests AI could affect complaints handling. Half of members said they have or would use AI to make a complaint, and 64% would not declare it, potentially increasing the volume and complexity of cases for schemes.
Martin Davidson, LPPA's chief technology officer, said members ‘want reassurance that oversight and accountability remain at the heart of important decisions.’
Burges Salmon partner Gary Delderfield said administering authorities should consider how to protect members and funds from AI-related risks.
