Fraud arms race ramps up
IDnow sealed a partnership to deliver continuous risk intelligence with Trustfull as Experian flags a fraud paradox—losses rose even while firms deployed AI—so data quality and adaptive detection are becoming table stakes. Meanwhile, real‑time deepfake voice detection tools like Modulate’s Velma are being pitched as "always‑on" protections for claims and SIU workflows. (biometricupdate.com) (artificialintelligence-news.com) (efficientlyconnected.com)
The IDnow Trust Platform centralises identity verification, behavioural signals and device-level telemetry to correlate identity, behavioural, device and digital data in real time for lifecycle risk decisions, and the release lists eIDAS 2.0, AMLR and PSD3 as supported compliance frameworks. (idnow.io) Trustfull’s announcement names concrete signal types it will feed into that platform—email, phone, device, IP and browser data—to detect disposable credentials, device spoofing and proxy usage across onboarding and subsequent sessions. (trustfull.com) IDnow’s insurance vertical explicitly calls out use cases across the policy lifecycle—onboarding policyholders, processing claims and verifying beneficiaries—positioning continuous identity signals as inputs insurers already map to underwriting and SIU triage. (idnow.group) Experian’s January 13, 2026 Future of Fraud Forecast reports 64% of businesses in EMEA & APAC saw rising fraud losses and 68% said existing fraud tooling cannot keep pace with AI-enabled attacks, with 73% planning to invest in device and behavioural data. (experianplc.com) Experian’s wider reporting and prior 2025 survey quantify the scale: nearly 60% of companies reported year‑on‑year increases in fraud losses from 2024 to 2025, FTC data cited $12.5 billion in U.S. consumer fraud losses in 2024, and Experian states its prevention solutions helped clients avoid an estimated $19 billion in global fraud losses. (experianplc.com) (businesswire.com) Modulate’s March 31, 2026 Velma Deepfake Detect launch names specific technical claims—ranked #1 on the Hugging Face Deepfake Speech leaderboard, built on an Ensemble Listening Model, priced from $0.25 per hour of audio and framed as delivering up to 578x lower cost than the next-best model to enable full-call monitoring. (modulate.ai) Industry commentary highlights operational changes enabled by Velma-style APIs: real-time scoring, streaming detection and integration into call-routing and escalation workflows make “always-on” voice validation feasible for contact centers, helpdesks and claims-handling pipelines that previously sampled only small portions of calls. (efficientlyconnected.com)
