Notifications
Clear all
Topic starter
15/08/2026 6:13 pm
TL;DR: AI-powered attacks now account for 41% of attacks targeting organisations, and Fingerprint’s survey of more than 300 leaders in SaaS, fintech, payment platforms, and banking shows average annual losses of $414,000, with more than a third reporting losses above $1 million. Privacy-first tools and regulations are also making legitimate users harder to distinguish from fraudsters, so fraud teams need stronger device intelligence and faster risk-based decisioning.
NHIMG editorial — based on content published by Fingerprint: State of AI Fraud & Privacy Report highlights
By the numbers:
- 41% of attacks targeting organisations are now AI-powered.
- 27% describe the impact of privacy-first technologies on, fraud prevention as severe.
- B2B SaaS fraud teams report a 62% rise, in manual workload from AI-powered attacks.
Questions worth separating out
Q: How should fraud teams adapt controls when AI-powered attacks scale faster than review capacity?
A: They should move from static rule sets to layered risk decisions that combine device intelligence, behaviour, and transaction context.
Q: Why do privacy-first technologies make fraud prevention harder?
A: Because they reduce the continuity signals fraud teams use to recognise returning devices, sessions, and users.
Q: What do teams get wrong about device intelligence in fraud prevention?
A: They often treat it as a standalone detector instead of an enrichment layer.
Practitioner guidance
- Rebalance fraud signals for privacy-constrained environments Review which device, session, and behavioural signals still remain reliable when users adopt privacy-first browsers or VPNs.
- Separate onboarding trust from ongoing trust Use different controls for account creation, login, and transaction approval so AI-generated abuse cannot rely on one verification decision to carry the whole lifecycle.
- Track analyst backlog as a security metric Measure manual queue growth, median investigation time, and repeat-case volume as indicators that fraud controls are failing to classify new attack patterns.
What's in the full report
Fingerprint's full report covers the operational detail this post intentionally leaves for the source:
- Sector-by-sector survey breakdowns for B2B SaaS, fintech, payment platforms, and banking.
- The report's wider data set on AI-powered fraud losses and investigation burden.
- Practical discussion of device intelligence as a fraud signal before login or account creation.
- The article's full framing on how privacy regulations are reshaping detection and user verification.
👉 Read Fingerprint's full State of AI Fraud & Privacy Report →
AI-powered fraud and privacy tools: what should fraud teams change?
Explore further
15/08/2026 6:20 pm
AI-powered fraud is now an identity-verification problem, not only a fraud-scoring problem. The article shows that attackers are using AI to increase scale, adaptiveness, and plausibility, which weakens controls that depend on static patterns or one-time verification. For identity teams, the important shift is that trust now has to be continuously re-evaluated across the user journey, not just at onboarding.
A question worth separating out:
Q: Who is accountable when privacy controls reduce fraud detection accuracy?
A: Accountability should sit with the teams that own both identity risk and customer experience, because the trade-off is operational as well as technical. Security, fraud, privacy, and product stakeholders need a shared policy for what evidence is required before blocking, challenging, or passing a user.
👉 Read our full editorial: AI-powered fraud is driving higher losses and heavier manual triage
