Security firm Microblink’s report Mapping the Rise of AI-Powered Identity Fraud describes how AI’s widespread availability and ease of use have changed how criminals commit fraud, and how security companies stop them.
AI has erased human limitations. It is now easy to operate across multiple methods, including text, voice, documents and video. This allows criminals to take more complex approaches and at scales previously unimagined.
In 2027, American fraud losses due to GenAI are estimated to hit $40 billion. 30 per cent of enterprises will consider ID verification methods unreliable due to deepfakes. Deepfake document forgery and related AI-driven identity fraud rose sevenfold since 2024.
While richer economies are still popular targets, AI makes it feasible to swim downstream, too. That brings the benefits of competing against less experienced security teams and protection methods.
“In other words, the U.S. and Canada are targeted because they are lucrative,” the report states. “Smaller countries are targeted because they can present different verification challenges.”
Criminals adapt their methods by region, with Microblink’s report outlining three personas. In Europe and Asia, attackers have refined screen-based fraud. The tools needed to engage in these activities are commonplace.
North American attacks are more precise, as fraudsters work on the actual documents to manipulate different parts. The goal is for them to pass structural checks while becoming another identity. Detection relies on catching minor inconsistencies in image integrity, layering, and data relationships. More sophisticated documents beget more sophisticated fraud tools.
Many other regions still fall prey to older methods that can still bypass cruder checks.
“To a human reviewer or a basic system, everything checks out,” Microblink’s report explains. “But under deeper inspection, these documents betray themselves through missing depth, inconsistent light reflection, and material properties that don’t match genuine substrates. The illusion holds until you look at how the document behaves, not just how it looks.
“That’s where deeper validation starts to matter. Beneath the surface, we consistently find mismatches between what the document shows and what its underlying data claims. Barcodes, in particular, act like a quiet truth serum. When the encoded data doesn’t align with the visible information, it exposes a fundamental break in authenticity that surface-level checks miss.”
AI has also allowed fraud to progress from mainly occurring at isolated moments like checkout to constantly evolving across sessions, devices and channels. That allows synthetic identities to mature and accounts to be compromised long after the initial verification.
