This work presents a novel face‑deidentification system developed by the Computer Vision Laboratory of the University of Ljubljana (UL FRI). The system was showcased during the European Researchers Night 2024, where visitors could step in front of a camera and view their own and others images anonymised in real time. The approach replaces the visible facial region with a synthetically generated face while preserving the subjects pose, gestures, and selected attributes such as facial expression or accessories (e.g., glasses). Leveraging deep‑learning‑based image synthesis and facial attribute conditioning, the method generates anatomically plausible replacements that maintain the natural dynamics of the original video stream. The system demonstrates a practical, privacy‑preserving alternative to conventional masking or blurring techniques, offering robust identity obfuscation without sacrificing essential biometric cues required for downstream analysis. The demonstration underscores the feasibility of integrating artificial‑intelligence‑driven deidentification into public‑facing applications and highlights its potential for safeguarding personal data in real‑time visual interactions. More info: https://humanistika.si/en/dogodki/deidentifikacija-obrazov/