biometric identification
A procedure that automatically determines who you are based on the unique features of your body, such as your fingerprint, face, iris or voice.
In plain terms
You know the password, you carry the key with you, but the biometric credential is you yourself. Biometric identification is like when a doorman recognizes you by your face, only here a machine does the same, much faster and even among millions of people.
It is convenient because it cannot be forgotten or left at home, and is harder to guess or steal through fraud than a password. This is why your phone, your bank's app, and airport passport gates use it. However, it is not a secret: anyone can photograph your face, and you leave your fingerprints on objects.
However, it is different from other identifiers: you can change your password at any time, but you cannot change your fingerprint or your face. If such data is leaked, it is permanent, which is why data protection law (GDPR) treats it as special, highly protected data.
Artificial intelligence truly made it powerful: today's facial recognizers work on learned patterns and can identify people remotely, without your knowledge. This is why the topic became both a security and a civil liberties issue.
How it works
- Capture: a sensor (camera, fingerprint reader, microphone) records your biometric data.
- Template creation: the system generates a numerical sequence, a so-called biometric template, from the characteristic features of the recording and uses this for comparison. Many systems store only the template, but for example, the passport chip and official registers also store the image itself.
- Comparison: the new pattern is matched against stored templates—either one-to-one (yours alone) or one-to-many (against a full database).
- Decision: the system calculates a similarity score, and if it exceeds a set threshold, it signals a match; thus, the result is always an estimate, not certainty.
- Liveness check: better systems also verify whether a real, living person is present, rather than a photo, mask, or AI-forged video.
An everyday example
In the morning, you look at your phone and it unlocks because it recognizes your face. In the afternoon, you approve a transfer in a banking app with a fingerprint, and at the airport, an automated gate compares your face with the image stored in your passport chip. These are all one-to-one verifications. Identification in the narrower sense occurs, for example, when police search for an unknown face visible in surveillance camera footage within a database containing many thousands of photos.
Who is affected
A common misconception
Limits and risks
- Biometric data cannot be replaced: once it leaks, it is compromised for your entire life.
- Systems can be fooled by photos, masks, voice clones, or deepfake videos if there is no reliable liveness check.
- A false match can lead to an innocent person being suspected, so critical decisions must not rely solely on machine matching.
- It can operate remotely without your knowledge, enabling mass surveillance.
Similar, but not the same
- biometric authentication – Authentication verifies that you are indeed who you claim to be (one-to-one), while identification in the narrower sense searches among many people to determine who you are (one-to-many). Common language calls both biometric identification, but the law distinguishes them sharply.
- facial recognition – Facial recognition is only one method of biometric identification; alongside it are fingerprints, iris scans, voice, and gait.
- emotion recognition – Emotion recognition does not aim to find out who you are, but tries to infer from your face or voice what you are feeling.
Related concepts
biometria, identification, data protection, cybersecurity
facial recognition, fingerprint identification, iris recognition, voice identification, Remote biometric identification
biometric authentication, multi-factor authentication, deepfake, liveness check, Computer vision
A brief history
The scientific basis for using fingerprints for identification was established by Francis Galton's 1892 book 'Finger Prints', following work by several 19th-century pioneers. It reached the masses in 2013 when Apple introduced the fingerprint reader (Touch ID) in the iPhone 5s, followed by the facial recognition Face ID in the iPhone X in 2017.
Where things stand
As of late 2026, it is commonplace in phones, banking apps, and at border crossings; at the Schengen external borders, the EU Entry/Exit System (EES) also records the facial images and fingerprints of non-EU travelers. Since February 2025, the EU AI Act has generally prohibited law enforcement from remotely identifying people in real time in public spaces, with only a few narrow exceptions that require authorization. Meanwhile, because of deepfakes, liveness checks are becoming increasingly important.
In the news
- FICO: AI arms race is reshaping the fraud threat across EMEAOct 5, 2026 · Money
- Ireland's domestic AI legislation takes effect; Italy regulates police biometricsSep 16, 2026 · Jog
- Italy adopts AI Act implementing decrees on biometrics and employment lawAug 5, 2026 · Jog
Also known as
biometric identification · Biometric identity verification · biometria · biometric authentication · Remote biometric identification · biometric identification · biometrics · remote biometric identification
This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.