Topic starter
10/06/2026 10:33 pm
TL;DR: Emerging fraud patterns are being caught by an upgraded deepfake detector that uses instant online self-learning, according to SumSub, which reports that multi-step attacks rose 180% in 2025 to 28% of fraud detected on its platform globally. The shift shows why static model refresh cycles no longer match the speed of AI-driven fraud.
Editorial analysis by NHI Mgmt Group, based on content published by SumSub: “How Adaptive Deepfake Detection Revolutionizes Digital Fraud Prevention Approach”.
Key questions
Q: What breaks when deepfake detection relies on periodic model updates?
A: Periodic updates create a blind window between the emergence of a new fraud pattern and the system learning how to recognise it.
Q: Why do deepfakes change identity verification requirements?
A: Deepfakes change requirements because they let attackers create believable but false evidence that can pass weak visual checks.
A: Warning signs include more successful social engineering attempts, unusual identity verification failures, a rising volume of manual reviews, and fraud patterns that reappear after KYC is completed.
Practitioner guidance
- Audit model refresh latency Measure how long it takes new fraud patterns to move from detection to deployed control, then identify where the gap creates exposure.
- Correlate multiple verification signals Require identity proofing decisions to consider documents, device signals, geolocation, biometric liveness, and linked-account patterns together.
- Test for injection-based fraud paths Red-team the session layer, not just the media layer, so controls can catch injection methods that bypass visual inspection.
Bottom line: Deepfake fraud now evolves quickly enough that periodic detection refreshes can leave a practical gap between attack emergence and model response.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
This topic was modified 2 days ago by NHI Mgmt Group
11/06/2026 2:37 am
Static fraud models are now a governance liability, not just a technical limitation. The article shows that periodic deepfake updates leave a blind window in which new attack patterns can spread before the model is refreshed. That is a control design failure, not an accuracy problem. Risk teams should treat refresh latency as a first-class governance metric, because the attacker only needs one unprotected interval to win.
A few things that frame the scale:
- The average estimated time to remediate a leaked secret is 27 days, despite 75% of organisations expressing strong confidence in their secrets management capabilities, according to The State of Secrets in AppSec.
- Our research also found that organisations maintain an average of 6 distinct secrets manager instances, creating fragmentation that undermines centralised control.
A question worth separating out:
Q: Should organisations still keep human review in deepfake-heavy workflows?
A: Yes, but only where the workflow is high-risk and the model cannot adapt quickly enough on its own. Human review should be reserved for exception handling, escalation, and ambiguous cases, not as the main control for every transaction. Otherwise the review queue becomes the bottleneck while attackers exploit faster, automated paths.
👉 Read our full editorial: Adaptive deepfake detection exposes the limits of offline fraud models
12/06/2026 4:11 am
Static fraud models create a detection lag that deepfake attackers can exploit. The underlying failure is not that traditional models are useless, but that their update cycle assumes fraud patterns remain stable long enough to be learned safely. In an AI-driven fraud environment, that assumption no longer holds. The practical consequence is that verification teams must treat model staleness as an operational risk, not a maintenance detail.
A few things that frame the scale:
- Nearly 60% of companies reported that fraud losses were still increasing in 2025.
A question worth separating out:
Q: Should identity proofing teams prioritise live adaptation over scheduled retraining?
A: Yes, when fraud patterns change quickly. Scheduled retraining still matters for governance, but it cannot be the only defence against attacks that mutate within days or hours. Teams should prioritise controls that learn from live sessions while preserving oversight, testing, and escalation for high-risk decisions.
👉 Read our full editorial: Adaptive deepfake detection exposes the limits of offline fraud models
This post was modified 2 days ago by NHI Mgmt Group