Beware of financial scammers
Deepfake fraud is becoming an ongoing, multiyear corporate risk as synthetic voices continue to circulate undetected.
Deepfake-enabled fraud, which began as novel technological feats, is now an ongoing operational risk with a multi-year shelf life within the corporate ecosystem. According to deepfake-detection provider Resembl.AI, deepfakes typically remain in circulation for three and a half years.
Resemble.AI’s 2025 Deepfake Threat Report, published in March, references an incident in which a voice clone of the CEO of a German energy company remained in circulation for nearly six years, although it resulted in a loss of only €243,000 in 2019.
The damage caused by such attacks is difficult to determine; For the 41 documented incidents from the past year cited by the research, verified losses of only $74.9 million were recorded, with an average per-incident loss of $243,000. However, the authors noted that 71% of victims did not report financial losses, suggesting a higher amount of hidden liabilities.
“What makes them so effective is that they enable real-time impersonation and the integration of real and fake data to create synthetic identities,” said Dominic Forrest, CTO of biometric security vendor Iprov. “These are extremely difficult to detect, and once trusted, they can be used to bypass controls and commit fraud.”
AI Arms Race
Detecting deepfakes is a growing concern; The authors of the Resembl.AI report estimate that deepfake-based fraud attacks on corporations reached 8.5 billion potential incidents, ranging from audio impersonations of executives to doctored or fake images. Forrest said the most common targets are account opening, payment authorization, credential reset and high-value transactions.
Experts warn that telling deepfakes from genuine articles has become an AI-on-AI battle.
Generic AI models that produce deepfakes continually improve through scaling and data, while deepfake detectors rely on signals such as artifacts and anomalies, which disappear as the model is improved, said Siwei Liu, professor of computer science and engineering and director of the AI and Data Science Institute at the State University of New York at Buffalo.
“In practice, detectors lag by about six to 18 months on specific modalities,” he said. “But more importantly, they are pursuing a moving target whose failure modes are being actively optimized.”
Forrest suggests that companies move their identity verification away from a single check to a multi-layered approach: “You need to confirm that a real person is physically present, not a deepfake, while also analyzing the digital environment for signs of compromise. No signal alone should be trusted.”
This article was first published in the May edition Global Finance Journal.
