Shufti confirms a real person is in front of the camera, blocking printed photos, replayed video, 3D masks and AI-generated faces. Passive liveness runs from a single frame with nothing asked of the user, while 40+ ensemble models read skin texture, 3D depth and contour, plus synthetic patterns, injection and screen replay. Active liveness adds blink or head-turn prompts for extra assurance. iBeta Level 3 conformance, 0% false acceptance on spoofs, with a pass or fail returned by API.