AI-aided extraction of fingerprints from "peace sign" photos raises identity fraud risks in China
Chinese media reports warn that fingerprints can be extracted from high-resolution photographs of people making the "peace sign" gesture, potentially enabling identity fraud through AI deepfakes and voice synthesis. Experts advise avoiding the gesture in close-range photos and compressing images before posting to social media.
What happened
A Chinese anti-fraud television program demonstrated that fingerprints can be extracted from high-resolution photographs of people making the 'peace sign' gesture at close range. Fingerprints can be extracted with 100% success rate from photos taken within 1.5 meters where the hand is directly facing the camera; extraction probability is approximately 50% at 1.5 to 3 meters; and extraction becomes essentially impossible beyond 3 meters.
The program also demonstrated that facial information illegally captured from video calls can be misused by fraudsters through AI deepfake technology and voice synthesis to impersonate identity.
Reported impact
Experts warned that fingerprints are permanent and cannot be changed, and once leaked could pose long-term risks. The program highlighted the potential for identity fraud through the combination of biometric data extraction and AI-generated synthetic media.
Index analysis
The incident demonstrates the convergence of photography metadata extraction, biometric identification technology, and generative AI systems capable of creating convincing deepfakes. Ordinary social media practices—posting high-resolution photos with visible hands—create exploitable vectors for extracting biometric information.
Governance controls
Experts advised users to avoid making the peace sign gesture in photographs taken within close range, to compress image quality before posting to social media, to delete previously published high-resolution photos showing clear hand details, and to enable biometric and password double authentication. Users are also advised to avoid entering biometric information on unfamiliar devices.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.