A test of highly sophisticated biometric spoof attacks holds good news for BioID, which claims to be first in the world to have scored perfect accuracy in a Level C biometric presentation attack detection (PAD) evaluation, and the firm’s customers.
TÜV Informationstechnik GmbH (TÜVIT) carried out the ISO/IEC 30107-3-based test based on hyper-realistic silicon masks, AI-generated deepfakes and high-definition video replays, which the independent testing lab classifies as Level C attacks, according to the announcement. These kinds of attacks, BioID says, represent the cutting edge of real-world fraud attack threats. BioID is profiled in the 2025 Deepfake Detection Market Report & Buyer’s Guide from Biometric Update and Goode Intelligence as a deepfake protection technology provider.
BioID’s software successfully identified all spoof attempts, for a zero percent attack presentation classification error rate (APCER).
“This result isn’t just a celebration for BioID; it’s a milestone for the industry,” says BioID CTO Dr. Robert Frischholz. “With this Level C certification, we’ve demonstrated that biometrics can deliver the same ‘face-to-face’ human interaction without compromising security.”
The biometric PAD capability is built into the BioID Web Service (BWS), and therefore included in all BioID applications, whether deployed through SaaS, on-premise or in a private hosting environment to meet regulatory requirements, the company says.
BioID refreshed its compliance confirmation for Levels A and B of TÜVIT’s biometric PAD program based on ISO 30107 and the FIDO Biometric Certification Requirements v2.2 in July. The importance of refreshing PAD compliance evaluations was emphasized by biometrics testing experts interviewed for the 2025 Face Liveness Market Analysis & Buyer’s Guide, which also profiles BioID.
Level C is an extrapolation of presentation attack instrument (PAI) classification based on the ISO standard, but not present in it. PAI’s for Level C are defined in the FIDO Biometric Certification Requirements, and BixeLab tests for this level of biometric spoof attack.
