The Pensions Regulator (TPR) has warned trustees and scheme managers they need to act now to ensure the safe and responsible use of AI across the workplace pension sector.
This is TPR’s first formal guidance about the use of AI in the pensions industry. It says schemes need to establish clear governance structures around AI use, and ensure outsourced providers have similarly robust controls in place.
In order to achieve this TPR said schemes should carry out ongoing testing and monitoring of AI systems, regularly review risks and controls, develop clear data strategies to ensure compliance with data protection rules, and strengthen defences against AI-driven pension scams.
The regulator said it would publish more detailed guidance on this issue later this year, following industry engagement over the summer.
Publishing its initial plan the regulator said AI could significantly improve administration, member engagement and decision-making across pensions — but warned that trustees remain accountable for outcomes even where AI systems are used by third parties, such as advisers, administrators or service providers.
But TPR also highlighted growing risks linked to AI adoption, including AI-enabled scams, cyber threats and potential bias in automated decision-making.
The regulator added that schemes should seek professional advice “appropriately and proportionately” when implementing AI-related innovations.
TPR chief executive Nausicaa Delfas said maintaining trust would be critical as schemes increasingly adopt AI technologies.
She said: “AI has the potential to transform pensions for the better: improving how schemes are run, how members are supported, and how the system as a whole delivers value.
“But trust is the most valuable asset in our system, and that trust depends on the safe and responsible adoption of AI in members’ interests.”
This announcement comes as AI adoption accelerates across financial services, including pensions administration, member communications and fraud detection.
Alongside setting expectations for schemes, the regulator also outlined how it plans to expand its own use of AI and advanced analytics within supervision and enforcement work. This includes using AI-enabled tools to identify scams and remove fraudulent websites targeting pension savers.
According to TPR, an AI-supported process has already assessed more than 2,000 scam websites and enabled the removal of 29 high-risk sites.
The regulator said it is committed to publishing annual updates on progress as AI adoption across the pensions sector evolves.
Hymans Roberston head of digital strategy Scott Finnie says: “We’re pleased to see TPR providing guidance on AI and support the principles outlined. AI brings significant opportunity for better member outcomes whilst simultaneously opening up new risk vectors. Navigating the fast-moving landscape will require diligence, adaptation and strong collaboration among trustees, their industry partners and the regulator. An agreed ethical framework will aid trustees in the series of decisions they will need to make with the use of this developing technology.”
Lumera commercial director, data & dashboards Maurice Titley adds: “The Regulator’s AI plan, and their intention to publish more detailed guidance in 2026, is an important step in setting clear expectations for how AI should be deployed responsibly across the pensions industry.
“Their approach of engaging with industry is welcome as there are many areas where AI innovation is being considered, not least in response to some of the new demands that will flow from reforms in the Pensions Act, such as assigning members to default retirement pathways.
“As we enter a new era for the pensions sector in the UK, AI is set to be a critical driver of transformation in how providers achieve greater efficiencies, strengthen governance and improve the member experience.
“The Regulator is right to point out that AI will bring both opportunities and risks which will require the industry to centralise innovative operating models, human oversight and robust governance to achieve the confidence of trustees.
“Scalable technology and a clear data strategy will be essential in ensuring that automated decision making is working off accurate records to achieve good outcomes for members. With these data building blocks in base, providers will have the potential to improve administration, standards and outcomes right across the pensions sector, enhancing rather than replacing the expertise that defines the industry.”
