AI-103 Practice Question 48: Face Liveness Anti Spoofing
Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 5 · question 8 of 10
A neobank's mobile onboarding flow captures a live selfie to verify a new customer's identity against their uploaded ID photo. During testing, a tester defeats the check by holding a printed photo of someone else's face up to the camera, and the app accepts it.
What should the team add to close this gap?
Show answer and explanation
Answer:
Liveness detection exists to close exactly this gap: it "determines if a face in an input video stream is real (live) or fake (spoofed)," acting as "an important building block in a biometric authentication system to prevent imposters from gaining access to the system using a photograph, video, mask, or other means to impersonate another person." The documented architecture pairs a frontend SDK session with an app-server orchestrator that starts the session and retrieves the result, and liveness can be combined with face verification so the flow both confirms a live person is present and that the person matches the ID photo.
- A is wrong because a detection confidence threshold reflects how sure the service is that a face is present in the frame at all. A printed photo still contains a real, detectable face, so raising this threshold doesn't distinguish a live person from a picture of one. - B is wrong because face attributes describe properties of whatever face was detected: pose, occlusion, glasses. A printed photo held at the right angle can present the same attribute values as a live face, so attribute checks alone don't detect the spoof. - D is wrong because identification answers "whose face is this among many known people," not "is a live person present." It doesn't address why a photograph was accepted in the first place.
Reference: Face liveness detection