AI fraud surge: what practitioners need to act on now

TL;DR: AI-enabled attacks surged 1210% in 2025, with non-AI fraud also rising 195% at year end, while healthcare saw more than 15,000 unique bot calls and retailers recorded a 330% increase in AI fraud in two months, according to Pindrop research. The scale shift means fraud controls must move from static detection to channel-aware, adaptive identity governance.

Editorial analysis by NHI Mgmt Group, based on content published by Pindrop: “Inside the 2025 AI Fraud Spike”.

By the numbers:

• AI-enabled attacks surged 1210% in 2025, according to Pindrop.
• Non-AI fraud also increased 195% at the end of 2025, according to Pindrop.
• Among major retailers and Pindrop customers, AI fraud increased 330% in two months, according to Pindrop.

Key questions

Q: How should security teams respond when AI fraud starts overwhelming customer and support channels?

A: They should treat the problem as a workflow integrity issue, not only a detection issue.

Q: Why do AI-generated voice and synthetic media make account recovery more dangerous?

A: Because recovery workflows often accept trust once the claimant sounds or looks legitimate.

Q: What are the signs that a fraud program is falling behind bot-driven abuse and account takeover?

A: A program is falling behind when bot traffic rises, customer support becomes a frequent recovery path for compromised accounts, and fraud moves into peripheral channels such as loyalty, referrals, or discounts.

Practitioner guidance

• Build channel-specific fraud controls Separate voice, chatbot, web, and recovery workflows in policy so each channel gets different risk scoring, step-up checks, and response thresholds.
• Treat recovery as a privileged workflow Require stronger re-authentication, out-of-band confirmation, and evidence logging before password resets, refund issuance, or account changes.
• Instrument bot traffic as a fraud signal Correlate bot volume, request velocity, and repeated failed attempts with account state changes and support interactions.

Bottom line: AI fraud now exploits the same business workflows that resolve accounts, issue refunds, and reset access, which makes channel governance as important as detection.

Explore further

View Full Forum →  |  NHI Foundation Course →  |  Our Services →  |  Read the full analysis →