AI-driven fraud losses surged last year, research reveals industry inflection point
Pindrop's latest research shows that in 2025, AI-driven fraud attack volume surged 1210% year-over-year, with total losses against major US clients reaching approximately $1 billion. Generative AI and deepfake technology are shifting fraud from manual operations to automated supply chains, with retail refund fraud being particularly prominent. A concurrent Trustpair survey also found that 71% of US enterprises experienced an increase in AI fraud.
Key Takeaways
- Fraud detection company Pindrop's recent research found that AI-driven attacks against its major U.S. clients surged by 1,210% last year, with total estimated losses reaching $1 billion.
- A report released this month noted that generative AI and deepfake tools are helping fraudsters scale attacks and improve their effectiveness. A Pindrop spokesperson said the research analyzed internal fraud data from more than 50 of its major U.S. clients.
- Pindrop CEO Vijay Balasubramaniyan said in an email response that last year "was a watershed moment for fraud shifting from manual operations to automated supply chains." He noted: "AI fraud is faster, cheaper, harder to detect, and is precisely targeting real-time decision-making scenarios: contact centers, urgent payment requests, and executive impersonation. Attackers can now use synthetic identities that look and sound authentic to launch thousands of attack attempts simultaneously across voice, video, chat, and email channels."
Deeper Insights
Atlanta-based Pindrop provides deepfake detection tools to identify scenarios such as fraudsters impersonating job applicants. The company noted that AI fraud has surpassed traditional fraud, marking a significant shift in the threat landscape.
The report stated: "AI-generated videos and voices can impersonate job applicants to gain system access, or impersonate executives to carry out social engineering scams. The specific methods vary, but the underlying logic is the same: convincing and highly sophisticated AI-driven scams."
This research follows similar findings from Paris-based Trustpair, a software vendor focused on payment fraud prevention.
Trustpair's research showed that over the past year, nearly three-quarters (71%) of U.S. businesses experienced an increase in AI-driven fraud attempts, and about half (47%) of finance leaders said AI-generated fraud has become one of the biggest challenges in fraud prevention. Among surveyed companies, one-quarter reported six-figure losses due to fraud.
Trustpair CEO Baptiste Collot said in a statement: "AI has raised the bar for fraud. Risks continue to rise, but the response speed of internal processes is far from sufficient."
Pindrop noted that retail was one of the industries hardest hit by AI fraud last year, with fraudsters particularly focused on scaling automated refund scams.
The company stated in the report: "Attackers deploy bots carrying common scripts to initiate return requests on retail websites. This strategy has a clear intent: targeting small refunds below a certain threshold to avoid triggering suspicion."