By Rob Sheehan, Head of Product
In the world of identity fraud, biometric deepfakes get a lot of attention. The exponential increase in biometric fraud and the power of GenAI to impersonate a face, fingertips or voice is both frightening and fascinating. Readily available tools to create hyper-realistic images and video or swap one face for another to disguise digital appearance have escalated awareness of the risk of biometric cloaking.
Although these threats and concerns are legitimate and must be counteracted, the arms race of fraudsters vs remote identity proofing is not confined to biometrics. Criminal document presentation attacks are increasingly common, both AI-generated ID forgeries and stolen images of legitimate passports and ID cards.
A genuine identity document is the anchor point of a secure online identity verification. The successful, secure authentication of that document is of first-order importance to protect against fraud, before checking for biometric presentation attacks that attempt to fake the genuine presence of the owner of that ID.
Document-based presentation attacks are increasing with the use of GenAI and other tampering techniques to manipulate identity documents
AI deepfakes, synthetic identities and document fraud
In recent years, identity document fraud has rapidly evolved from crude forgeries to an industrialized, AI-enabled business pushing out high-resolution counterfeits. Criminal ‘ID-as-a-service’ marketplaces now sell ready-made ID cards, passports, document templates and onboarding kits on the dark web. High-resolution microprinting and sophisticated forgery of document security features, including hologram and UV laminate has enhanced replication. Document supply chain compromises have also enabled large batches of genuine documents to be abused.
Cybersecurity data breaches continue to release vast amounts of genuine personal data (names, dates of birth, addresses, Social Security Numbers) into this black market for fake ID. The criminal practice of ‘document morphing’ —blending real and fake identity data—stitches together plausible, synthetic identities from real and fabricated attributes into the pages of a fake passport or printed onto a phoney driving licence.
Document presentation attacks are increasing
The result is the increasing use of AI and GAN (Generative Adversarial Network) models to manipulate identity documents and verification. Document-based presentation attacks are increasing. The rise in the sophistication of AI technologies is giving criminals a platform to commit identity fraud; the number of fake documents used for identity fraud targeting business is having a global impact.
In the UK, Cifas 2025 Fraudscape research found that 30% of ‘false applications’ were due to counterfeit documents, especially targeted at sectors where identity document checks are routinely required, such as banking and the financial services.
ID-Detect prevents document presentation attacks
ID-Detect is an advanced document fraud detection solution that adds an AI-driven document authentication layer to ID-Pal’s existing verification technologies, confirming that the document presented is genuine, original and in hand.
ID-Detect instantly creates a list of forensic checks for the identity document image submitted by an applicant at onboarding, looking for evidence of fraud. These checks
examine both the document image itself, and the identity data captured from the image to check that it is authentic.
By checking for and eliminating the most common types of document fraud, ID-Detect proves a genuine, physical ID is present and detects advanced document presentation attacks.
From the most sophisticated to the simplest forms of fraud, ID-Detect prevents identity document presentation attacks.
AI deepfake documents and digital manipulation
AI-generated forgeries and criminal digital manipulation of ID in document presentation attacks are both increasingly common. ID-Detect’s AI-driven document authentication engine spots evidence of digital manipulation and the markers of presentation attacks. Our models are trained to detect discrepancies like pixelation, texture differences and pattern irregularities and general inconsistencies, ensuring that screen-based fraud is flagged immediately and diverted away from your onboarding flow.
A suspicious document flagged by ID-Detect is essentially “quarantined” with a clear reason code, giving your team detailed insights for faster, more confident decisioning. With reduced false positives, a smoother onboarding experience for genuine customers, and flexible configuration options that adapt to your business needs, ID-Detect empowers you to protect against AI-driven fraud without creating friction for trusted users.
The result of this advanced ID fraud detection is that 98% of fake documents are stopped automatically while genuine
customers are rapidly onboarded Independent testing of one fraud detection check deployed by ID-Detect was found to achieve over 99% effectiveness in identifying the fraudulent documents it was designed to uncover.*
Document photo substitution attacks
A portrait swap, also known as ‘portrait paste over’, can be ether a physical or digital substitution of the legitimate photograph displayed on an identity document. This common identity document fraud substitutes the genuine photo on an ID with another so that the document may appear authentic but shows the wrong person.
ID-Detect spots texture differences and pattern consistencies between the document and the overlaid portrait, to tell whether the document has been tampered with. We analyse the edges of document portraits for any shading or other irregularities that might signal fraud.
Document screen replay attacks
A screen replay of an ID is a common document presentation attack. In this simple fraud technique, criminals exploit the remote identity verification process. Instead of capturing and uploading a real-time image of a physical identity document, a pre-recorded screenshot or video of an ID is replayed on a device screen in order to fool the system. And the identity document they capture may have even been digitally manipulated or AI-generated to begin with. ID-Detect counters this by detecting Moiré patterns, caused by interference between screen pixels and the camera sensor. We also look out for pixelation, backlight inconsistencies, and reflection detection, picking up unnatural reflections or glares that are common when images are displayed on screens ensuring that screen-based fraud is flagged early in the process.
Printed copy attacks
High‑quality scans or photographs of genuine IDs are another common form of document presentation attack. Beyond colour inconsistencies, we use checks to identify resolution degradation, printer noise artefacts, and gloss.
Authentic IDs use specific laminations and materials, which consumer printers cannot replicate, Analysing the surface reflectivity and texture allows our system to reject counterfeit documents with high accuracy.
The result of these combined ID fraud detection checks is that fake identity documents are caught automatically and removed from the onboarding flow, while genuine customers are rapidly signed up.
Winning the arms race against fake IDs and synthetic identity fraud
Every ID-Detect pass and fail decision automatically generates audit-ready evidence on a tamper-proof log entry, complete with a record of captured ID document images and process timestamp. These results are ready in seconds, so your compliance team and regulators can refer to verification decisions.
With every attempted identity document presentation attack, the algorithms that drive ID-Detect document authentication continue to learn. Universal document coverage allows you to instantly verify IDs from anywhere in the world without extra setup, costly retraining, or the risk of gaps in coverage. Paired with facial matching biometric technology and secure liveness testing, ID-Pal ensures that fraudsters must attempt to beat several ID security defences at once, while genuine customers may be onboarded to your business in moments.
ID-Pal document security fitted as standard
Delivered as a standard security feature of all ID-Pal plans, ID-Detect delivers an AI-driven document authentication layer to ID-Pal’s multilayered verification platform. To find out more about defending your business from identity document presentation attacks, from the simplest to the most sophisticated criminal tactics, get in touch with the ID-Pal team.
* Result for testing of Printed Copy presentation attacks.
