AFSecurity at UiO
Amina Bassit from Mobai AS gives a talk on the challenges of reusable identity with privacy-preserving biometrics.
TIME: 9 October 2026, at 14:00h
PLACE: Seminar room Java (2423), 2nd floor, IFI, UiO
All interested are welcome. Coffee and snacks served.
AGENDA:
14:00h Welcome to AFSecurity at UiO
14:05h Invited talk
- TITLE: Challenges of reusable identity with privacy-preserving biometrics
- SPEAKER: Amina Bassit, biometrics and cryptography lead at Mobai AS
- ABSTRACT: A passport is a trusted source: cryptography proves the document is authentic, and a photo binds it to its holder. Online, that binding is checked not by a border guard but by a face recognition system with attack detection, and its quality sets the assurance level of the whole identity-proofing flow. Unlike on-device biometrics such as Face ID, a reusable identity requires the proofing result to outlive the session and to be relied on later, by other parties and from other devices. The reference template must therefore be stored, yet a face template is invertible, linkable and permanent, so nobody may ever see it. This talk shows how homomorphic encryption makes that possible: enrolment and matching run on encrypted data, and the comparison preserves accuracy under a standard cryptographic security argument. The decisive questions then become where the encrypted reference lives and who holds the key. Starting from a threat model with the person, the service provider and the matcher as parties, the talk compares the available cryptographic approaches and argues that whether the parties are fixed or change over time determines which one fits. It closes with what a protected reference lets a service do and the attack surface that every reuse reopens.
14:45h Discussion.
| BIO: Amina Bassit is the Secure Biometrics and Cryptography Lead at Mobai AS, where she leads the innovation research work on privacy-preserving reusable identity, using homomorphic encryption to perform biometric recognition without exposing the underlying biometric data. She holds a PhD from the University of Twente on biometric search under encryption, which received the EAB Industry Award in 2024, and was previously a postdoctoral researcher in privacy-preserving machine learning at Michigan State University. Her research has led to several patents, and she has presented tutorials on biometric privacy and security at IJCB 2024 and on computer vision over homomorphically encrypted data at CVPR 2025. |   |
| AFSecurity seminars on cybersecurity are organised by the UiO Digital Security Research Group. |  |