TransUnion Strengthens Device Security with New Machine Learning Capabilities

WHY THIS MATTERS

Digital fraud is escalating at a pace that is outstripping traditional detection methods, with businesses reporting losses of over $500 billion globally. The sharp rise in account takeovers and increasingly sophisticated attack methods—such as the use of virtual machines, proxies and identity masking—has made legacy, rule-based fraud systems less effective. TransUnion’s expansion of machine learning capabilities within its Device Risk solution reflects the growing need for adaptive, intelligence-driven fraud prevention that can evolve alongside attacker tactics.

Suspected digital fraud continues to impact businesses worldwide. In a recent TransUnion (NYSE: TRU) survey of 1,200 business leaders, respondents reported fraud losses totaling $534 billion. To help companies combat this growing threat, TransUnion today announced expanded machine learning (ML) capabilities within its Device Risk solution.

The enhancements are designed to help organizations detect and combat increasingly sophisticated attacks, while maintaining a streamlined and trusted customer experience. Today’s announcement comes at the Merchant Risk Council’s MRC 2026 conference in Las Vegas, where TransUnion will be exhibiting its fraud solutions at Booth 422.

To help businesses stay ahead of emerging threats, TransUnion Device Risk has been further powered to enable:

• Stronger recognition of returning devices across customers
• More robust detection of non‑human activity (including behavior patterns associated with virtual machines, residential proxies and remote desktops)
• Deeper consortium-driven insights that illuminate evolving fraud trends

These updates enhance fraud‑detection accuracy and streamline digital customer experiences by reducing unnecessary friction. The new capabilities introduce advanced machine learning that extends Device Risk intelligence far beyond traditional, static rule-based decisioning.

Pre-built adaptive ML models learn from thousands of device signals and fraud feedback sourced from TransUnion’s long-standing global fraud consortium. This enables proactive detection of anomalies and evasion attempts. ML has demonstrated the ability to improve fraud capture by up to 50%, while also reducing the volume and complexity of manually maintained rules, lowering operational overhead, and improving overall precision.

Steve Yin, global head of fraud at TransUnion said: “Traditional device fingerprinting has been impacted by privacy-driven technology changes and evolving tactics that let fraudsters look like ‘new’ users with just a few clicks. We need to meet this moment with solutions that learn continuously, adapt in real time, and connect more signals across more browsers and applications. This will enable more effective recognition of risky behavior even as identifiers change. These enhancements mark a significant advancement in how device-level intelligence is used to secure digital interactions across industries”.

Digital Fraud Rising Across the Globe

Digital fraud continues to expand across the global economy. According to TransUnion’s H2 2025 Update to its Top Fraud Trends Report, organizations lost an average of 7.7% of equivalent annual revenue to fraud over the past year. At the same time, the volume of digital account takeovers and the rate of suspected digital fraud tied to account creation both increased over the last year.

According to a TransUnion analysis, suspected digital account takeovers increased by 141% between H1 2024 and H1 2025. Over the same period, suspected digital fraud at account creation grew by 26%. These trends underscore the need for more adaptive, precise and intelligent device recognition capabilities.

“Our Device Risk enhancements demonstrate how TransUnion innovates to stay a step ahead of advanced fraud tactics by pairing richer device‑level intelligence with adaptive machine learning,” said Clint Lowry, vice president of global fraud solutions at TransUnion. “By elevating both detection and efficiency, we empower customers to operate with greater confidence across login, transaction and account creation experiences.”

FF NEWS TAKE
Fraud prevention is shifting from static rules to dynamic, data-driven intelligence.

TransUnion’s enhancements highlight how machine learning is becoming essential for detecting modern fraud patterns that constantly evolve. As digital interactions increase across industries, the organisations that invest in adaptive, real-time fraud detection will be better positioned to reduce losses while maintaining seamless customer experiences.