evidoria

← Back to browse

Good practice Imported

AI-Powered Scam-Site Detection — The Pensions Regulator and PSAG's Machine-Learning Tool Against Pension Fraud

United Kingdom · Brighton · See the United Kingdom profile

The UK's Pensions Regulator and Pension Scams Action Group built a machine-learning tool to flag fraudulent pension websites: 830 sites reviewed, 29 high-risk ones taken down, 94 referred to partner agencies, against £17.7m lost to pension scams in 2023.

AI-Powered Scam-Site Detection — The Pensions Regulator and PSAG's Machine-Learning Tool Against Pension Fraud

Details

Promoter
The Pensions Regulator (TPR) and the Pension Scams Action Group (PSAG)
Period
Tool unveiled April 2025; in active use as of mid-2026
Keywords
fraud detection, pension scams, machine learning, financial regulation, consumer protection

Description

The Pensions Regulator (TPR), the UK's pensions watchdog, worked with the multi-agency Pension Scams Action Group (PSAG) to build a machine-learning tool that scans the web for websites impersonating or facilitating fraudulent pension schemes. The tool was unveiled at an industry webinar in April 2025 attended by more than 550 pensions-industry participants.
As of the tool's public reporting, PSAG had reviewed 830 websites, had 29 assessed as high-risk taken down, and referred 94 to partner agencies for further action. TPR's own research found that only 11% of pension schemes correctly knew to report suspected scams to Action Fraud, the UK's national fraud reporting service, despite existing guidance - one of the gaps the new tool is meant to help close.
The context for the tool is substantial financial harm: according to Action Fraud, more than £17.7 million was lost to pension scams in 2023, at an average loss of £46,959 per victim. A parallel initiative, a new Fraud and Cyber Crime Reporting and Analytics Service, is due to replace Action Fraud later in 2025 to improve national fraud-intelligence sharing with police.
Public reporting to date covers detection and takedown activity but not downstream outcomes - there is no published figure yet for money recovered or scams prevented as a direct result of sites the tool flagged, and the 29-of-830 takedown rate reflects an early-stage, low-volume operation rather than a mature, audited enforcement programme.

Read the full analysis: https://www.thepensionsregulator.gov.uk/en/media-hub/press-releases/2025-press-releases/pension-scam-fighters-stand-united-as-artificial-intelligence-takes-the-fight-to-the-fraudsters

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful