Summary of "Face Liveness Market Overview"

Purpose and scope

A joint market report and buyers’ guide on face liveness (also known as presentation attack detection, PAD).
Covers market drivers, adoption, standards and testing, vendor landscape, forecasts, and practical buyer guidance.
Authors / hosts: Good Intelligence (Alan Good, CEO & chief analyst) in partnership with Biometric Update (Chris Burt, managing editor).

Key findings and market positioning

Face liveness has evolved from a sub‑component of face matching to a standalone product category focused on answering both:
- “Is this a real human?” and
- “Is this the right identity?”
Four primary adoption drivers:
1. Fraud prevention (including AI‑driven attacks)
2. Enhanced security
3. Reduced user friction (passive/hybrid methods)
4. Improved biometric accuracy
Leading verticals: financial services, government, enterprise, gaming & gambling (including age checks), e‑commerce/retail, healthcare, and travel/border control.
Primary applications: remote digital onboarding, biometric authentication (step‑up auth, account recovery), document authentication (portrait substitution detection), and account management (password reset/device setup).

Forecast headline: face liveness market revenue is projected to exceed about $252 million by 2027. (The report includes three‑year forecasts for transactions and revenue.)

Standards, testing and independent assurance

Standards matter: ISO and other regional/industry standards are referenced (the transcript also mentions a “phto alliance” standard).
Independent testing and continuous R&D/testing are emphasized as essential for vendor claims.
Good Intelligence interviewed multiple independent test labs to understand testing types, limitations, and how buyers should interpret results.
Labs named in the presentation (transcribed): IBAA (US), a lab in France (“F”), Ingenium (UK), and Bixie Lab (Australia).
Labs advised buyers to ask vendors about:
- production performance (post‑lab),
- operational best practices,
- data management,
- continuous monitoring and model improvement,
- incident/risk mitigation processes.
Note: PAD testing differs from NIST‑style FVT. Vendors that engage labs for compliance testing typically have mature solutions and expect good outcomes; PAD testing workflows and costs differ from face matcher evaluations.

Vendor landscape and the “20 pioneers”

The report highlights 20 pioneering providers. The transcribed list (auto‑generated subtitles may contain errors) includes:
- aware, bioid, CyberLink, dayon, fa Tech, ID R&D, idemia, Erse, incode, iProov, jumio, Mobi, onfido, paravision, Ro, socure, trinamix, veridos, y, Erse
Presentation cautioned that companies vary in methods, specializations and appropriate use cases — buyers should match vendor strengths to their specific use case.
Buyers are encouraged to consider many suppliers because the market is crowded and specialized.

What to look for in a face liveness provider — 7 baseline criteria (buyer checklist)

Cost
- Does pricing meet budget and expected transaction volumes? (Expect regional/sector variance.)
Accuracy & reliability
- Independent testing/certification, ability to detect deepfakes/morphs and injection attacks, beyond minimum standards.
Security posture
- Cyber certifications, secure SDLC, vulnerability programs (bug/specialty bounties), SOC capabilities.
Usability / UX fit
- Active vs passive vs hybrid liveness options, latency, user friction and mobile experience.
Privacy & compliance
- GDPR and local biometric data laws; data storage/processing governance.
Integration
- Ease of embedding into existing workflows, SDK/API support, support for account management and step‑up flows.
Ongoing operations & monitoring
- Ability to update models, monitor accuracy in production, incident response and risk mitigation.

Technical & operational topics covered

Liveness modalities
- Active, passive and hybrid approaches. Passive/hybrid are emphasized for frictionless UX.
Deepfakes, morphs and injection attacks
- Deepfake and morph detection is a major focus.
- Injection attacks (bypassing PAD by injecting media) are treated as distinct cyber attacks and require additional countermeasures.
- The boundary between PAD and broader cybersecurity is blurring.
Software vs hardware
- Most remote/mobile PAD will be software‑based.
- Hardware solutions exist (e.g., Trinamix / device‑embedded 3D) but are less flexible for broad remote deployment.
- Note: Apple’s 3D face tech was noted as not being a sellable liveness product.
SaaS vs embedded
- SaaS/PaaS models offer faster update cycles, centralized visibility into attack patterns, model improvements, and the ability to operate SOCs or collect telemetry for threat intelligence.
- Embedded/on‑device solutions can lag in update speed (dependence on OS vendors).
Continuous improvement mechanisms
- Vendors use security/bug/spoof bounty programs, novel testing, internal SOCs and threat intelligence feeds to stay ahead in the arms race with attackers.

Practical recommendations & buyer questions to ask vendors

Key questions:
- How does the solution perform in production (post‑lab)?
- How do you handle personal data and ensure compliance?
- How do you monitor and measure accuracy in real operations?
- How are models updated and how frequently?
- What incident risk mitigation and response processes exist?
Validate certifications and independent test results; check whether tests measure aspects relevant to your use case (e.g., deepfake/injection resilience vs standard PA attacks).
Treat liveness/PAD as part of your cybersecurity and fraud defense stack, not just an identity add‑on.

Materials and availability

Attendees were told a recording and a 10% discount link for the report would be emailed to registrants.

Main speakers and sources

Chris Burt — Managing Editor, Biometric Update (host/presenter)
Alan (Allan) Good — CEO & Chief Analyst, Good Intelligence (report co‑author, lead analyst)
Independent testing labs mentioned (transcribed): IBAA (US), a French lab (“F”), Ingenium (UK), Bixie Lab (Australia)
Report produced by Good Intelligence in partnership with Biometric Update; features a list of 20 pioneering vendors (see Vendor landscape). Note that the transcribed vendor list may contain subtitle errors.