Shufti forecasts 39X more deepfake ID documents attacks this year

Deepfakes of identity documents are exploding, and there will be 39-times more attacks with deepfaked IDs this year than last, according to a new report from Shufti, rendering selfie biometrics checks performed in isolation obsolete.

Overall deepfake fraud is expected to surge nearly five-fold, growing by 495 percent from last year to this, according to Shufti’s Deepfake Fraud Index.

Synthetic identity attacks outnumbered live video deepfakes by 50 percent, according to Shufti’s figures, making up more than 42 percent of total AI fraud. Face swaps made up just 17.6 percent, and document deepfakes 11.9 percent. But that balance is about to change.

Shufti’s estimate that ID document deepfakes will spike by 3,900 percent is based on annualizing fraud rates from January to May of this year.

The 2025 Deepfake Detection Market Report and Buyers Guide from Biometric Update and Goode Intelligence forecasts face deepfake detections will grow from just over 1.4 million last year to more than 5 million in 2027.

The report from Shufti breaks down attacks by type, and explores the layered combination of checks that must complement selfie biometrics to make them operationally reliable. These include liveness detection, injection attack detection and media forensics software, along with four other measures Shufti takes as part of its defense against deepfakes.

Tests of deepfake detection technology in laboratory settings can help customers identify potential candidate deepfake detection providers, but Shufti CTO Frayam Asif argued that “(t)he industry needs to stop treating lab accuracy as deployment readiness” in a recent company blog post.

“The conditions under which we verify identity bear almost no resemblance to the conditions under which we test for fraud,” Asif explained.

Article Topics

deepfake detection  |  deepfakes  |  document liveness  |  market report  |  Shufti  |  synthetic identity fraud