The future of identity operations is here. Meet Persona AI →
Published November 20, 2026
Last updated November 20, 2026

Persona achieves highest CEN/TS 18099 certification level for injection attack detection

Independent lab testing confirms Persona holds up against the most demanding injection attacks, from rooted devices to tampered camera services.
Irfan Faizullabhoy
Irfan Faizullabhoy
4 min
Persona achieves CEN/TS 18099 High certification for injection attack detection
Key takeaways
Injection attacks: An attacker replaces the camera feed with a deepfake or stolen video, and the system receives a seemingly live capture with nearly no visual artifacts.
CEN/TS 18099 High: Persona earned the highest certification for catching injection attacks after CLR Labs evaluated its Android SDK.
Regulatory expectations: Regulators are starting to require injection attack testing.

Today Persona earned the highest certification for catching injection attacks: CEN/TS 18099 High. Cabinet Louis Reynaud (CLR) Labs, an ISO/IEC 17025-accredited lab that helped write the first ever standard for testing injection attack detection, evaluated Persona’s Android SDK.

Persona’s defenses against injection attacks are fully automated and already run in real time across millions of verifications a day.

“Our customers are facing more sophisticated deepfakes, face-swaps, and replay attacks, while identity verification volumes are growing rapidly. Earning CEN/TS 18099 High is one way we’re helping them stay ahead of both, with independently tested defenses that work at scale.”

Irfan Faizullabhoy, Product Lead at Persona

Injection attacks expose a gap in traditional liveness testing

For years, the industry’s go-to standard for liveness testing has been ISO/IEC 30107-3, which primarily inspects what is displayed to a real camera: a printed photo, a mask, a screen. What’s missing from that standard is protection against injection attacks, where an attacker replaces the camera feed with a deepfake or stolen video instead of showing something to the camera, and the system receives a seemingly live capture with nearly no visual artifacts.

In a Gartner survey published in September 2025, 62% of organizations said they had faced a deepfake attack in the previous 12 months.

“While deepfake submissions continue to grow in volume and sophistication, an even more alarming trend is emerging with agentic AI. AI is making it easier to replicate legitimate customer experiences and we’re observing increased real time phishing attacks in which bad actors mimic verification flows, trick users into submitting government IDs and selfies, and attempt to inject that genuine media into legitimate verification sessions. This makes detecting and preventing injection attacks more critical than ever.”

Daniel Lee, Product Lead at Persona

How CLR Labs tested injection attack detection

CEN/TS 18099, published by the European Committee for Standardization (CEN) in 2024, is the first standard for testing injection attack detection. Evaluators first try to break in, using tools like emulators and hooked apps. Then they use that access to deliver fake media, such as deepfakes and manipulated images. Results are graded on three levels, based on how much time and expertise an attack requires. Each level covers the attacks below it.

  • Basic. Attacks that take little time or expertise.

  • Substantial. Attacks that take more effort, such as feeding a virtual webcam or emulator into the capture flow.

  • High. The most demanding attacks, such as rooting a device, tampering with its camera service or kernel, or running a custom Android build to bypass the camera.

Evaluators check that genuine users still get through, so strictness alone doesn’t earn a pass.

“The Cabinet Louis Reynaud, through its conformity certification department CLR CERT® congratulates Persona on achieving High-level compliance against the CLR Cert IAD certification program under CEN/TS 18099. As editor of the European standard on biometric injection attacks, we built this evaluation to test defenses against injected video, tampered apps, and modified operating systems. In a constantly changing digital world, compliance is more essential than ever. We are honored that Persona has selected CLR Cert.”

Kévin Carta, Director of CLR Labs

Governments and standards bodies are also responding to these threats, updating their expectations for how identity verification systems should defend against them.

New standards are shaping regulatory expectations for identity verification

Regulators are starting to require injection attack testing.

  • EU. ETSI TS 119 461 v2.1.1, the standard for remote identity proofing by trust service providers, requires CEN/TS 18099 testing. It sets High as the bar for its Extended level, which qualified trust service providers (QTSPs) must meet. The EU Anti-Money Laundering Regulation, which covers remote customer identification, applies from July 10, 2027.

  • US. NIST’s revised Digital Identity Guidelines (SP 800-63-4) set requirements for defending against injected media.

  • International. ISO is developing ISO/IEC 25456 with CEN/TS 18099 as its starting point. Publication is planned for 2027.

To add certified injection attack detection to your verification flow, contact Persona.

The information provided is not intended to constitute legal advice; all information provided is for general informational purposes only and may not constitute the most up-to-date information. Any links to other third-party websites are only for the convenience of the reader.

Irfan Faizullabhoy
Irfan Faizullabhoy
Irfan Faizullabhoy is a product manager at Persona. He wakes up early to surf in Half Moon Bay and likes to warm up with a jazz record and coffee.

Continue reading