The short answer
Short answer: Yes, fingerprint ridges can be visible in suitable high-resolution photographs. AI can enhance information already captured in an image. However, whether it can produce an accurate fingerprint suitable for an attack depends on many factors and has not been generally demonstrated for ordinary selfies.
Why a victory-sign selfie matters
When someone makes a victory sign, their fingertips often face the camera. In good lighting and with a sharp image, a camera may record very fine ridges on the skin. Parts of the individual's ridge pattern may be visible depending on image quality, angle and distance.
In May 2026, the South China Morning Post reported on a demonstration on Chinese television. Financial expert Li Chang showed how image processing and AI could make the skin ridges in a celebrity photograph more visible. He pointed to conditions involving relatively short camera distance and fingertips facing the lens.
That demonstration illustrates a technical possibility. It does not prove that any arbitrary selfie reliably reveals a complete and authentic fingerprint. In particular, reported distances should not be treated as universally validated security thresholds.
Is this really a new threat created by AI?
No. Security researchers have studied the reconstruction of biometric features from photographs for many years.
As early as 2014, Jan Krissler, known as “Starbug”, demonstrated at the Chaos Communication Congress the risks high-resolution photographs can pose to biometric authentication. Today's AI discussion adds more powerful image-enhancement tools to an older problem.
Modern image processing can increase contrast, reduce noise or plausibly fill in detail. An important distinction follows: a line that looks convincing after sharpening does not necessarily match the person's real skin pattern. Looking clear is not the same as being reconstructed accurately.
Could someone unlock my phone with it?
Not merely because the photograph exists.
An attacker would first need sufficiently accurate biometric features. They would then have to reproduce these in a form accepted by the particular fingerprint sensor. They would also need suitable access to the device or authentication process.
Cisco Talos demonstrated in April 2020 that such protections are not inherently impossible to bypass. In a controlled experiment, researchers fooled different fingerprint sensors using specially manufactured artificial fingerprints. This required several steps and considerable effort.
The reported success rates applied to the selected sensors and attack procedures in that experiment. They cannot be extrapolated to random social-media photographs.
Why biometric data is different
A password can be replaced. Your own fingerprint is, in principle, persistent. Biometric authentication therefore has a different security profile from a secret password.
The US National Institute of Updatedards and Technology (NIST) states an important principle: biometric characteristics are not secrets. Authentication systems using them need additional technical safeguards, including protection against presented replicas, restrictions on failed attempts and suitable additional authentication factors.
This does not mean fingerprint sensors are generally unsafe in everyday use. Risk depends on the device, its protection mechanisms, the attacker and the particular attack scenario.
What can individuals and businesses do?
Take a closer look at fingertips in photographs shared publicly. Especially with very high-resolution close-ups, avoid unnecessary visibility of fine skin ridges. Cropping, reducing detail or deliberately obscuring sensitive areas may help.
Use strong access controls for phones and accounts. Choose a secure PIN or strong password and use available additional security features. Biometrics can make authentication convenient, but should not be treated as a secret or as replaceable in the same way as a password.
Businesses should pay attention too. Photos from events, trade-fair stands and workplaces can unintentionally show access passes, screens, confidential records and, under favourable conditions, biometric details. A brief review before publication makes sense.
The point is not to prohibit victory-sign selfies. It is to understand the technical possibilities and assess real-world risks proportionately.
Frequently asked questions
Can AI read a fingerprint from every selfie?
No. The picture must contain enough usable detail. Sharpness, resolution, lighting and camera angle matter greatly. AI cannot guarantee correct recovery of authentic features that were never captured.
Does a photographed fingerprint automatically make a phone insecure?
No. A photograph alone cannot unlock a device. A successful attack requires further steps and depends heavily on the sensor and its safeguards.
Should people stop posting photographs with outstretched fingers?
A blanket ban is unnecessary. For high-resolution, publicly accessible close-ups, review sensitive details and reduce them if appropriate.
Conclusion
AI can reveal how much information seemingly harmless photographs may contain. Under suitable conditions, fingerprint ridges can be reconstructed to some extent. This does not automatically produce a reliably identical fingerprint or enable a successful phone attack.
For individuals and businesses alike, a considered approach to sharing high-resolution images is the most practical response.
Want to learn how AI can be used in business processes in a controlled, auditable way? Explore more articles in TechniKInside or take a look at my services.
Explore services ↗Sources and further reading
- South China Morning Post, 10.05.2026, report on the Chinese demonstration: https://www.scmp.com/news/people-culture/trending-china/article/3352777/china-tv-variety-show-exposes-scam-linking-peace-sign-selfies-privacy-risks
- Chaos Computer Club, 31C3, Jan Krissler, presentation, 27 December 2014: https://media.ccc.de/v/31c3_-_6450_-_de_-_saal_1_-_201412272030_-_ich_sehe_also_bin_ich_du_-_starbug
- Cisco Talos, „Fingerprint cloning: Myth or reality?“, 08.04.2020: https://blog.talosintelligence.com/fingerprint-research/
- NIST, Digital Identity Guidelines, SP 800-63B, Abschnitt
Use of Biometrics: https://pages.nist.gov/800-63-4/sp800-63b.html