Gartner®, Deepfake Identity Threats: Mitigate Risk in Identity Verification and Face Biometrics, Akif Khan, Nayara Sangiorgio, James Hoover, 11 May 2026
Gartner® is a trademark of Gartner, Inc. and/or its affiliates.
Share
Share
Key Takeaways
- Liveness detection is the first line of defense in biometric identity verification, confirming a real, present person before any face match, document validation, or data checks run.
- CLEAR1 uses PAD2 (ISO/IEC 30107-3)–certified, passive liveness detection to help stop presentation attacks such as printed photos, video replays, masks, and AI-generated deepfakes—without forcing users through clunky head-turn or blink challenges.
- Instead of treating liveness as a standalone gate, CLEAR1 combines this signal with document authenticity checks, source validation, and device intelligence to build a multi-layered view of identity.
- With CLEAR1’s reusable identity model, liveness checks don’t just secure onboarding—they support fast, secure reverification for account recovery, privileged access, and other high-risk workflows across the user lifecycle.
Why Liveness Detection Matters for Modern Identity
Liveness detection—one of the most important parts of modern identity verification—answers a critical question: is there a real, present person behind this interaction, or is the system looking at a spoofed image, replayed video, or AI-generated deepfake?
The answer to that question is only the starting point. To keep up with deepfakes, synthetic identities, and compromised devices, liveness detection has to sit inside a broader, multi-layered approach to identity.
That’s where CLEAR1 comes in. CLEAR1 layers liveness with biometric comparison, document checks, device security signals, and source validation to move beyond basic verification to true identity assurance.
What is Liveness Detection?
Liveness detection is the process of determining whether a biometric sample—typically a face image captured from a selfie or camera feed—comes from a real, live person present at the time of capture rather than from a spoofed or synthetic source.
In a typical identity verification flow, a user takes a selfie on a mobile device or computer. Before that selfie is compared against an ID photo, liveness detection analyzes the image and related signals to decide: is this a real human face with natural depth, texture, and lighting, or is it a presentation attack—a printed photo, screen replay, mask, or injected video?
Once liveness has been confirmed, the system can move on to biometric comparison and the rest of the identity checks, with the confidence that there is a real, present person behind the image.
What Liveness Detection Actually Does
Generative AI has dramatically reduced the cost of creating convincing fake identities. Synthetic personas, high-resolution printed or screen-based spoofs, and replayed video can all be engineered to bypass basic visual checks. At the same time, attackers are increasingly focused on impersonating legitimate users—not just breaking technical controls.
In that context, liveness detection has to reliably distinguish a real, present human from printed or screen-displayed photos, 2D and 3D masks, replayed or injected video streams, and AI-generated faces and deepfakes tuned specifically to fool biometric systems.
And yet, even a perfect liveness signal is not enough to verify an individual if the document is forged, the device is compromised, or the underlying data doesn’t tie back to a real person. That’s why CLEAR1 treats liveness as the opening move in a multi-layered identity strategy.
Active vs. passive liveness (and why it matters)
Most discussions about liveness detection focus on a trade-off between security in active liveness and passive liveness.
Active liveness asks users to perform actions such as turn their head, blink, smile, follow a dot with their eyes. This can be effective, but it adds friction—and can even lead to terminating the experience on slower devices or in less-than-perfect network conditions.
Passive liveness runs in the background. All the user has to do is take a selfie while neural networks and computer vision models analyze that single frame (or a short burst of frames).
Often, the difference between these two kinds of liveness is viewed as a choice between security (active) and convenience (passive). CLEAR1 challenges that framing by using passive liveness detection that has been independently certified at PAD Level 2 under ISO/IEC 30107-3—a standard designed to validate biometric systems against a broad spectrum of presentation attacks.
With CLEAR1’s advanced liveness protection, organizations can maximize security while minimizing friction.
What Powers Liveness Detection Software
While implementations vary by vendor, modern liveness detection software typically evaluates a combination of:
- Skin texture and micro-patterns that distinguish real skin from printed or screen-displayed surfaces
- Face geometry and ratios that should remain consistent across different poses and lighting
- Light and shadow behavior across the face, which reveals whether the image has true depth or is a flat surface
- Image metadata and capture characteristics, which can expose anomalies in how and when the image was produced
Neural networks trained on large datasets of both genuine and attack images learn to spot artifacts such as:
- Moiré patterns and glare from photographing screens
- Unnatural edge behavior around facial features
- Texture inconsistencies in masks, printed photos, or AI-generated faces
All of these signals are combined into a single decision, making liveness a fast behind-the-scenes check that turns a complex defense into a simple yes-or-no answer.
What PAD2 certification actually means
ISO/IEC 30107-3 defines how biometric systems should be tested against presentation attacks. Within that framework, Presentation Attack Detection (PAD) Level 2 is a critical benchmark. PAD2 certification indicates that an independent lab has validated a system’s ability to resist a wide range of known attack species in controlled testing, including:
- Printed photos and display attacks
- 2D and 3D masks
- Prosthetics and wax heads
- Video replay and injected streams
- AI-generated deepfakes and other advanced spoofs
CLEAR1’s liveness detection is PAD2-certified, meaning the system has been externally tested against this broader attack surface, not just self-attested to follow the standard.
How Liveness Detection Works in the CLEAR1 Flow
- Capture a selfie
The user takes a selfie in a guided flow that checks for sufficient image quality, lighting, and framing.
- Run passive liveness detection
CLEAR1’s PAD2-certified liveness detection models evaluate the selfie to confirm there is a real, present person, not a spoofed or synthetic image. If liveness fails, the process stops.
- Compare the live selfie to the ID photo
Once liveness is confirmed, biometric matching compares the selfie to the portrait on a government-issued ID to determine whether the same person appears to be presenting both.
- Validate the document and corroborate identity data
In parallel, CLEAR1 analyzes the ID itself for authenticity—checking security features, layout, and chip-level data where supported—and corroborates key identity attributes that are extracted from the ID across authoritative and credible data sources.
- Assess device security signals
Device and network signals help surface tampering, emulation, or other suspicious behavior tied to the endpoint, phone number, or email—signals that are increasingly important as attackers look for ways around traditional controls.
- Reach an identity decision
By this point, liveness is one of hundreds of real-time signals—not a standalone check—feeding into CLEAR1’s identity assurance model to verify the person beyond the device.
Why Businesses Choose CLEAR1
For enterprises, liveness detection is not just a feature box to check—it’s a core part of balancing fraud reduction, user experience, compliance, and long-term trust. CLEAR1 is built to do exactly that. Here’s how:
PAD2-certified, passive liveness
CLEAR1’s liveness detection has been independently validated at PAD Level 2 under ISO/IEC 30107-3, helping organizations defend against advanced presentation attacks while letting users complete identity verification with a single selfie.
Multi-layered identity assurance
Liveness is one layer in a broader multi-layered approach that draws on biometrics, document authenticity, device security, and source validation to confirm users are who they claim to be.
Reusable identity and a large existing network
Over 43 million people in the CLEAR network can verify instantly with just a selfie, and new users gain the same reusable identity after a quick, one-time setup. That means the strongest, most thorough checks happen once. After that, liveness-backed reverification is intuitive and fast.
Enterprise-grade compliance
CLEAR1’s verification stack—including liveness detection—supports IAL2, AAL2, HIPAA, SOC 2, and PAD2 requirements, helping organizations align with demanding regulatory and security standards.
Continuous accuracy and bias testing
CLEAR1 continuously evaluates False Match Rates (FMR) and False Non-Match Rates (FNMR) across demographics to help minimize bias and keep performance consistent over time
Proven performance and ROI
Across customers, CLEAR1 has helped reduce fraud, deliver verification that is significantly faster than traditional methods, and automate account recovery at scale—cutting account-related support calls by more than half.
Brand trust that accelerates adoption
89% of people associate CLEAR with security and trust—an advantage that helps drive user adoption of security controls like liveness-backed verification.
Where businesses deploy liveness detection
CLEAR1 runs liveness as a standard check in every verification, so enterprises rely on it across multiple touchpoints, including:
- Customer onboarding and KYC, to help ensure new accounts are tied to real people, not synthetic identities or deepfake-driven impersonation attempts.
- Account recovery and password or MFA resets, where liveness-backed selfies can replace knowledge-based questions that are increasingly easy for attackers to answer.
- Workforce access and IAM workflows, where verifying the person beyond the device helps prevent impersonation, social engineering, and insider-style threats.
- Patient check-in and portal access in healthcare, where liveness helps confirm the person behind sensitive health data, prescriptions, and billing information.
- Ongoing reverification at high-risk actions, such as privileged access changes, data exports, or large financial transactions.
Confirm the Person, Not Just the Document
Liveness detection shifts the question from “does this credential look right?” to “is there a real person here who matches this identity?” CLEAR1 utilizes that answer as one signal in a multi-layered approach to identity verification, blending biometrics, documents, devices, and real-world data to help stop fraud.
See how CLEAR1 goes beyond a basic liveness check to verify identity across every critical touchpoint.
Frequently Asked Questions
What is liveness detection?
Liveness detection is a biometric security control that determines whether the system is interacting with a real, physically present person at the moment of capture, rather than a spoofed image, replayed video, or deepfake. In CLEAR1’s flow, a liveness check runs on the user’s selfie before any biometric match, document verification, or source corroboration takes place.
What is the difference between active and passive liveness detection?
Active liveness detection prompts users to take specific actions—like turning their head, smiling, or following an on-screen target—to prove they are present. Passive liveness detection analyzes a single selfie (or a short burst of frames) in the background, without extra prompts. CLEAR1 uses passive liveness detection, combining a low-friction experience with PAD Level 2 certification to maintain strong resistance against presentation attacks.
How does liveness detection stop deepfakes?
Liveness detection models are trained to recognize artifacts and inconsistencies that often appear in deepfakes, such as unnatural texture patterns, irregular lighting and shadows, and subtle distortions around facial features. When combined with PAD2-grade testing, these models help block AI-generated faces and replayed videos before they ever reach the biometric comparison step.
What does PAD2 certification mean for liveness detection?
PAD2 (Presentation Attack Detection Level 2) is an independently validated benchmark under ISO/IEC 30107-3. It confirms that a liveness system has been tested against a wide range of presentation attacks—including printed photos, masks, video replays, injected streams, and deepfakes—and has met defined thresholds for detecting them. CLEAR1’s PAD2-certified liveness detection gives enterprises confidence that their passive liveness check has been externally vetted, not just self-certified.
Where in an identity verification workflow does liveness detection run?
In CLEAR1’s architecture, liveness detection runs at the very start of the identity verification flow, immediately after selfie capture and before any biometric comparison, document authenticity check, or source validation. If liveness cannot be confirmed, the workflow can be stopped or escalated, preventing downstream systems from making decisions based on spoofed or synthetic inputs.
Is liveness detection the same as facial recognition?
No. Liveness detection asks whether there is a real, present person in front of the camera. Facial recognition (the biometric comparison step) asks whether that person matches a known or claimed identity—such as the portrait on a government-issued ID. CLEAR1 separates these stages deliberately: a user must pass a liveness check first, then a biometric match, and then additional document, device, and data checks as part of a broader identity assurance strategy.
Why Liveness Detection Matters for Modern Identity
Liveness detection—one of the most important parts of modern identity verification—answers a critical question: is there a real, present person behind this interaction, or is the system looking at a spoofed image, replayed video, or AI-generated deepfake?
The answer to that question is only the starting point. To keep up with deepfakes, synthetic identities, and compromised devices, liveness detection has to sit inside a broader, multi-layered approach to identity.
That’s where CLEAR1 comes in. CLEAR1 layers liveness with biometric comparison, document checks, device security signals, and source validation to move beyond basic verification to true identity assurance.
What is Liveness Detection?
Liveness detection is the process of determining whether a biometric sample—typically a face image captured from a selfie or camera feed—comes from a real, live person present at the time of capture rather than from a spoofed or synthetic source.
In a typical identity verification flow, a user takes a selfie on a mobile device or computer. Before that selfie is compared against an ID photo, liveness detection analyzes the image and related signals to decide: is this a real human face with natural depth, texture, and lighting, or is it a presentation attack—a printed photo, screen replay, mask, or injected video?
Once liveness has been confirmed, the system can move on to biometric comparison and the rest of the identity checks, with the confidence that there is a real, present person behind the image.
What Liveness Detection Actually Does
Generative AI has dramatically reduced the cost of creating convincing fake identities. Synthetic personas, high-resolution printed or screen-based spoofs, and replayed video can all be engineered to bypass basic visual checks. At the same time, attackers are increasingly focused on impersonating legitimate users—not just breaking technical controls.
In that context, liveness detection has to reliably distinguish a real, present human from printed or screen-displayed photos, 2D and 3D masks, replayed or injected video streams, and AI-generated faces and deepfakes tuned specifically to fool biometric systems.
And yet, even a perfect liveness signal is not enough to verify an individual if the document is forged, the device is compromised, or the underlying data doesn’t tie back to a real person. That’s why CLEAR1 treats liveness as the opening move in a multi-layered identity strategy.
Active vs. passive liveness (and why it matters)
Most discussions about liveness detection focus on a trade-off between security in active liveness and passive liveness.
Active liveness asks users to perform actions such as turn their head, blink, smile, follow a dot with their eyes. This can be effective, but it adds friction—and can even lead to terminating the experience on slower devices or in less-than-perfect network conditions.
Passive liveness runs in the background. All the user has to do is take a selfie while neural networks and computer vision models analyze that single frame (or a short burst of frames).
Often, the difference between these two kinds of liveness is viewed as a choice between security (active) and convenience (passive). CLEAR1 challenges that framing by using passive liveness detection that has been independently certified at PAD Level 2 under ISO/IEC 30107-3—a standard designed to validate biometric systems against a broad spectrum of presentation attacks.
With CLEAR1’s advanced liveness protection, organizations can maximize security while minimizing friction.
What Powers Liveness Detection Software
While implementations vary by vendor, modern liveness detection software typically evaluates a combination of:
- Skin texture and micro-patterns that distinguish real skin from printed or screen-displayed surfaces
- Face geometry and ratios that should remain consistent across different poses and lighting
- Light and shadow behavior across the face, which reveals whether the image has true depth or is a flat surface
- Image metadata and capture characteristics, which can expose anomalies in how and when the image was produced
Neural networks trained on large datasets of both genuine and attack images learn to spot artifacts such as:
- Moiré patterns and glare from photographing screens
- Unnatural edge behavior around facial features
- Texture inconsistencies in masks, printed photos, or AI-generated faces
All of these signals are combined into a single decision, making liveness a fast behind-the-scenes check that turns a complex defense into a simple yes-or-no answer.
What PAD2 certification actually means
ISO/IEC 30107-3 defines how biometric systems should be tested against presentation attacks. Within that framework, Presentation Attack Detection (PAD) Level 2 is a critical benchmark. PAD2 certification indicates that an independent lab has validated a system’s ability to resist a wide range of known attack species in controlled testing, including:
- Printed photos and display attacks
- 2D and 3D masks
- Prosthetics and wax heads
- Video replay and injected streams
- AI-generated deepfakes and other advanced spoofs
CLEAR1’s liveness detection is PAD2-certified, meaning the system has been externally tested against this broader attack surface, not just self-attested to follow the standard.
How Liveness Detection Works in the CLEAR1 Flow
- Capture a selfie
The user takes a selfie in a guided flow that checks for sufficient image quality, lighting, and framing.
- Run passive liveness detection
CLEAR1’s PAD2-certified liveness detection models evaluate the selfie to confirm there is a real, present person, not a spoofed or synthetic image. If liveness fails, the process stops.
- Compare the live selfie to the ID photo
Once liveness is confirmed, biometric matching compares the selfie to the portrait on a government-issued ID to determine whether the same person appears to be presenting both.
- Validate the document and corroborate identity data
In parallel, CLEAR1 analyzes the ID itself for authenticity—checking security features, layout, and chip-level data where supported—and corroborates key identity attributes that are extracted from the ID across authoritative and credible data sources.
- Assess device security signals
Device and network signals help surface tampering, emulation, or other suspicious behavior tied to the endpoint, phone number, or email—signals that are increasingly important as attackers look for ways around traditional controls.
- Reach an identity decision
By this point, liveness is one of hundreds of real-time signals—not a standalone check—feeding into CLEAR1’s identity assurance model to verify the person beyond the device.
Why Businesses Choose CLEAR1
For enterprises, liveness detection is not just a feature box to check—it’s a core part of balancing fraud reduction, user experience, compliance, and long-term trust. CLEAR1 is built to do exactly that. Here’s how:
PAD2-certified, passive liveness
CLEAR1’s liveness detection has been independently validated at PAD Level 2 under ISO/IEC 30107-3, helping organizations defend against advanced presentation attacks while letting users complete identity verification with a single selfie.
Multi-layered identity assurance
Liveness is one layer in a broader multi-layered approach that draws on biometrics, document authenticity, device security, and source validation to confirm users are who they claim to be.
Reusable identity and a large existing network
Over 43 million people in the CLEAR network can verify instantly with just a selfie, and new users gain the same reusable identity after a quick, one-time setup. That means the strongest, most thorough checks happen once. After that, liveness-backed reverification is intuitive and fast.
Enterprise-grade compliance
CLEAR1’s verification stack—including liveness detection—supports IAL2, AAL2, HIPAA, SOC 2, and PAD2 requirements, helping organizations align with demanding regulatory and security standards.
Continuous accuracy and bias testing
CLEAR1 continuously evaluates False Match Rates (FMR) and False Non-Match Rates (FNMR) across demographics to help minimize bias and keep performance consistent over time
Proven performance and ROI
Across customers, CLEAR1 has helped reduce fraud, deliver verification that is significantly faster than traditional methods, and automate account recovery at scale—cutting account-related support calls by more than half.
Brand trust that accelerates adoption
89% of people associate CLEAR with security and trust—an advantage that helps drive user adoption of security controls like liveness-backed verification.
Where businesses deploy liveness detection
CLEAR1 runs liveness as a standard check in every verification, so enterprises rely on it across multiple touchpoints, including:
- Customer onboarding and KYC, to help ensure new accounts are tied to real people, not synthetic identities or deepfake-driven impersonation attempts.
- Account recovery and password or MFA resets, where liveness-backed selfies can replace knowledge-based questions that are increasingly easy for attackers to answer.
- Workforce access and IAM workflows, where verifying the person beyond the device helps prevent impersonation, social engineering, and insider-style threats.
- Patient check-in and portal access in healthcare, where liveness helps confirm the person behind sensitive health data, prescriptions, and billing information.
- Ongoing reverification at high-risk actions, such as privileged access changes, data exports, or large financial transactions.
Confirm the Person, Not Just the Document
Liveness detection shifts the question from “does this credential look right?” to “is there a real person here who matches this identity?” CLEAR1 utilizes that answer as one signal in a multi-layered approach to identity verification, blending biometrics, documents, devices, and real-world data to help stop fraud.
See how CLEAR1 goes beyond a basic liveness check to verify identity across every critical touchpoint.
