Deepfake identity fraud surges as criminals go professional
Fri, 20th Nov 2026 (Today)
Entrust has published its 2027 Identity Fraud Report, which found that sophisticated identity fraud rose sharply in 2026.
The findings point to a split in the fraud landscape. Overall document fraud fell 19% over the year, but sophisticated document fraud rose 1,870% and advanced document fraud increased 268% as criminals shifted to harder-to-detect methods.
This suggests many organisations have become better at stopping basic forms of identity fraud, including simple alterations and template-based attacks. In response, fraud groups appear to be using more skilled operators, better source material and more complex tactics.
"The big story is that fraud has become a professional operation," said Simon Horswell, Senior Fraud Specialist Manager at Entrust.
"Fraudsters are investing in better templates, stolen credentials and specialized skills to make attacks more convincing. AI is accelerating this shift, helping fraudsters scale operations and target multiple points across the identity lifecycle."
Deepfake growth
The report recorded a 165% rise in deepfake attacks year on year, equal to one detected deepfake every three minutes in 2026. Financial services accounted for 94% of all deepfakes detected globally, while attacks against digital-first banks rose 714%.
Banking fraud is also shifting towards existing customers. Account takeovers represented 52% of banking fraud, compared with 48% for new account fraud, indicating that fraudsters are increasingly targeting live accounts rather than only trying to open new ones.
A compromised financial account can be used for money laundering, money muling and other follow-on crimes, making successful identity fraud more valuable to organised criminal networks.
"Deepfake tools are getting faster, easier to use and more capable, which will improve the quality and realism of their output," Horswell said.
"That raises the stakes for financial institutions, where a successfully passed identity check can open the door to far more than one fraudulent transaction. Banks need to deploy layered defenses that can assess the authenticity of identity media alongside the person, device and behavior across each interaction."
Injection attacks
Another area of sharp growth was injection fraud. Injection attack volume increased 224% in 2026, outpacing deepfake growth.
These attacks involve feeding images or video directly into verification systems rather than presenting them through a live camera on a phone or laptop. Attackers use tools such as virtual cameras and device emulators and, in some cases, inject genuine identity media and biometric data obtained through scams, coercion or data breaches.
Device intelligence fraud detections rose 42%, underlining the need to examine device and behavioural signals as well as the identity media itself. The report argues that genuine stolen content can still be used fraudulently if a system relies too heavily on a single check.
"Injection attacks show why layered defenses are critical," Horswell said.
"While the injected content itself may be authentic, real-time monitoring of device and session behavior can reveal signs of fraud."
Human targets
The report also found that reported social engineering attempts rose 520% in 2026 and increased 75-fold over two years. It described social engineering as the fastest-growing attack method in its dataset.
That trend reflects a shift towards targeting people with legitimate access, including customers and employees who can approve transactions, recover accounts or complete security procedures. Once criminals obtain valid credentials through manipulation, they can use trusted channels to take over accounts.
Regional patterns
Entrust's regional data showed marked differences in how fraud is developing. The Americas had the highest fraud rate, at 6%, and the highest share of physical counterfeit documents, at 68%. The region also recorded a 2,200% rise in deepfake attacks, the fastest annual growth rate in the report.
Europe remained the largest centre for deepfake activity by share, accounting for 61% of attacks recorded in the study. It also had the highest share of sophisticated identity fraud attempts, at 43%.
In Asia-Pacific, the highest level of basic identity fraud was 74%, while digital forgeries made up 47% of identity fraud. Deepfake activity in the region rose 116% year on year.
The report is based on analysis of identity verification data collected between September 2025 and September 2026. Entrust said its identity verification systems process millions of checks each year across 195 countries and that the findings were normalised by client and industry distribution.
"Fraud is professional, multichannel and increasingly focused on both new and existing accounts," said Tony Ball, Chief Executive Officer at Entrust.
"Organisations must protect identities across their own operations, customer interactions and third-party providers, especially at critical moments such as account opening, account recovery, authentication and high-risk transactions. Identity verification, biometric authentication, device intelligence and trust signals work together to fight fraud today while providing a foundation for governing AI agent actions. The future of identity security is moving beyond one-time checks toward continuous trust throughout the identity lifecycle."