The Pensions Regulator (TPR) in the UK has published a new AI Plan, outlining its expectations for the responsible use of artificial intelligence (AI) in workplace pensions and warning that trustees and scheme managers must act now to protect savers from emerging AI-driven risks.

The regulator said AI had the potential to “transform pensions for the better” by improving administration, member engagement and decision-making, but stressed that accountability for outcomes would remain with trustees and scheme managers, even where AI systems are used.

As part of the plan, TPR confirmed that it will publish more detailed guidance later in 2026 following engagement with the industry over the summer.

TPR highlighted several opportunities associated with AI adoption, including personalised member communications, improved fraud detection, faster administration processes and enhanced risk monitoring.

The regulator said schemes and providers were already using AI-powered tools to automate routine tasks and provide more tailored support to members.

However, the regulator also warned of growing risks linked to AI use, including AI-generated scams, increased cyber threats, bias in decision-making and concerns around members relying on unregulated AI tools for financial planning and retirement guidance.

TPR chief executive, Nausicaa Delfas, argued that AI could improve how schemes are run and how members are supported, but stressed that trust in the pensions system depended on safe and responsible adoption.

“Our message to trustees, administrators and scheme managers is clear: act now," she said.

"Put strong governance in place, invest in data quality, understand where and how AI is being used in your scheme, and protect your members from AI-driven fraud.”

Under the plan, TPR said trustees and scheme managers should establish clear governance and accountability structures for AI use, ensure robust testing and ongoing monitoring of AI systems, and regularly review risks and controls.

The regulator also urged schemes to improve data quality, maintain clear data strategies, and ensure compliance with data protection legislation on automated decision-making.

TPR outlined four priority areas for its own work on AI, including strengthening governance standards across schemes, improving data foundations, supporting responsible innovation, and using AI internally to become a more effective regulator.

As part of this, the regulator said it would continue working with the Financial Conduct Authority (FCA) to ensure regulatory alignment across the pensions sector and pension supply chain.

The regulator also confirmed that it has already been using AI-enabled processes to identify potential pension scam websites.

According to the plan, more than 2,000 websites have been assessed using the technology, leading to the removal of 29 high-risk sites while reducing manual scanning and triage work by around two hours a day.

TPR said it would report annually on progress relating to AI adoption and innovation within the pensions sector, including barriers to adoption and lessons learned.

This article originally appeared in our sister publication Pensions Age.

---