Skip to content
←Back to glossaryCyber

Liveness detection

A control meant to guarantee that a real, present human stands before the camera, the digital substitute for physical presence.

Definition

In August 2023, Hong Kong police broke up a ring that filed online loan applications with stolen identity cards and got past the lenders' facial recognition with AI-generated face swaps. During remote onboarding, liveness detection has to establish that the captured face belongs to a person who is actually there, and not to a replayed image. The market adopted it on a massive scale once remote onboarding became the norm, to replace physical presence at the counter. Deepfakes attack it head on (face swaps, virtual cameras, injection attacks), and the substitute turns out more fragile than what it replaced. Improving detection indefinitely is a race lost in advance. The defense is layering: passive and active liveness, injection detection, biometric binding to the document, continuous behavioral analysis.

Its 4 cross-references, woven across the glossary's 18 categories
Liveness detection: Cyber glossary illustration, AlgoPolis
Example

Textbook case: a security firm recorded more than eight thousand attempts to bypass the liveness controls of a single financial institution in a few months.

Read next
CyberDeepfakeA synthetic medium, voice or video, generated by artificial intelligence to convincingly imitate a real person, becoming a vector of fraud and disinformation.Read the entry →CyberInjection attackA technique that bypasses the camera outright by injecting a fabricated stream into the verification system.Read the entry →CyberSynthetic identityA person who never existed, assembled from real data (a Social Security number, a birth date) and invented details, whose fraud can go unnoticed for a long time because there is no victim to report it.Read the entry →
Put it to work
SimulatorPlay this case from the underwriter's seatEight quarters, one book, cyber and political risk files. You set the price and the war clause, then the market answers back.Take the seat →
Related terms
Stay informed

Follow the research

One letter a week. First, the week's underwriting note, when one is published: a dated, argued decision, signed by Paul Bernardski. Then what the observatory published on cyber risk, political risk and insurance. No account, no tracking.

No spam, one-click unsubscribe, no data ever sold.