Should Big Tech Help Cover the UK’s £1.3bn Fraud Losses?

Fraud losses in the United Kingdom reached £1.3 billion in 2025, intensifying debate over whether social media and technology companies should contribute to reimbursing scam victims.

In a recent industry discussion, UK Finance financial crime director Kathryn Westmore argued that online platforms should share responsibility. The episode also examined how criminals are adopting AI and other new technologies.

Scams span several sectors

A victim may first encounter a fraudulent advertisement on a platform, communicate through a messaging service and then send money through a bank or crypto provider. Controls concentrated at the payment stage therefore address only the final part of the chain.

Platforms can improve advertiser verification, remove fraudulent content and preserve data for investigations. Banks can use payment warnings, behavioural monitoring and beneficiary intelligence. Faster sharing between the two can stop repeat attacks.

The policy question is how to allocate liability without weakening incentives for any participant. Regardless of the final reimbursement model, firms should record scam origins and use that information to disrupt the channels that repeatedly generate losses.

Possible models for shared responsibility

Policymakers could allocate costs through mandatory contributions, case-by-case reimbursement or penalties where a platform failed to remove a known fraudulent advertiser. Each model creates different incentives and requires evidence showing where the scam originated.

Banks and platforms therefore need common data definitions and secure processes for exchanging campaign, account and device indicators. Victims should not have to navigate several organisations to prove the same event. A coordinated response can freeze beneficiary funds, remove malicious content and identify related attacks before losses spread.

Evidence needed for fair loss allocation

• Where and when the victim first encountered the scam.
• How the advertiser or account was verified by the platform.
• Warnings shown before the payment and the customer’s response.
• Whether earlier reports could have prevented repeat victimisation.

Next focus: Any liability model will need clear standards for causation, evidence preservation and appeals. Without them, disputes over responsibility could delay reimbursement and weaken incentives for rapid intervention.