The opinions expressed here are those of the authors. They do not necessarily reflect the views or positions of UK Finance or its members.

A growing threat to UK finance

Synthetic identity fraud is rapidly becoming one of the most damaging threats to the UK’s financial sector. Unlike traditional identity theft, synthetic fraud combines real and fake data—such as a valid National Insurance number and a made-up name—to create a new, seemingly legitimate profile.

These identities often pass as real customers: opening accounts, making transactions, and even building credit over time. But they’re designed to break out—defaulting on loans or maxing out credit before disappearing. In 2024, losses from synthetic identity fraud in the UK topped £300 million. And it’s still rising.

Onboarding is just the beginning

Fraud detection can’t stop at the application stage. Many synthetic profiles behave like genuine customers for months or years before executing an attack. Therefore, detection must be continuous, not one-off.

To stay ahead, financial institutions need tools that can surface hidden signals across the customer lifecycle. Jaywing’s machine learning development environment, Archetype, enables organisations to rapidly develop, train, and deploy tailored fraud models designed to detect synthetic patterns that static rules often miss. With Archetype, organisations can:

• Detect subtle anomalies across large datasets
• Score both new and existing accounts in real-time
• Adapt models in response to emerging threats
• Automate fraud detection while retaining transparency and compliance

This ability to evolve is critical in defending against sophisticated, slow-moving fraud.

A layered approach to detection

A strong synthetic identity strategy needs to span multiple touchpoints:

1. Identity verification at onboarding
Use biometric checks and real-time data to go beyond document validation and confirm applicants are real people, not constructed identities.

2. Device and behavioural analytics
Subtle signals like uniform typing speeds, lack of cursor movement, and repeated device use across identities can indicate non-human or scripted behaviour. Building behavioural and device profiles over time makes it easier to spot anomalies and flag suspicious activity.

3. Continuous monitoring
Many synthetic identities remain dormant. Monitoring transaction patterns, personal data changes and digital activity can catch fraud well after onboarding.

4. Machine learning risk scoring
Archetype can identify combinations of traits and behaviours that flag synthetic identities long after account creation.

5. Industry collaboration
Sharing intelligence is vital. Techniques like federated learning allow institutions to train models on shared insights without exposing sensitive data—strengthening fraud defences across the sector.

Fraud is no longer a single point of failure

Synthetic identity fraud isn’t a one-time breach. It’s a slow-build strategy. And that means fraud prevention must be just as persistent.

By combining behavioural analytics, advanced modelling, and sector collaboration, financial institutions can protect themselves against threats that don’t show up in day-one checks.

Fraud prevention needs to evolve. Because synthetic identities already have.