
The New Anatomy of Identity Fraud
How attackers exploit documents, images and digital identity signals — and where verification stops them.
5 min · October 2026
Read More: The New Anatomy of Identity FraudMissing a country or a custom check? Tell us what you need.Book a 15-min call
What is stored on the chip in a passport or ID card, how a phone reads and checks it, and why it tells you more than any photo of the document.
By secuveri teamPublished

A photo of an identity document shows what the document looks like. A good document check can learn a lot from that: whether the layout matches the issuer's template, whether the security features are where they should be, whether the check digits in the machine-readable zone add up. What it can't do is prove that the data on the card was written by the issuing authority.
Most passports, and a growing share of ID cards, carry a contactless chip that can. The chip holds the holder's data and portrait, signed by the issuing country. A smartphone with NFC can read it in a few seconds while the person holds the document against the back of the phone. This post explains what happens during that read and where its limits are.
The format is set by ICAO Doc 9303, the international standard for machine-readable travel documents. Passports have used it for many years, and identity cards issued in the EU since August 2021 must include a contactless chip under Regulation (EU) 2019/1157.
The chip stores data in numbered data groups. The two that matter most for verification are:
Alongside them sits the Document Security Object (SOD). It contains a hash of each data group, signed with a key that belongs to the issuing country. That signature is what makes the chip useful.
A chip does not hand its data to any reader that comes near it. The reader first has to prove it has seen the document. With Basic Access Control (BAC), the access keys are derived from the document number, date of birth and expiry date printed in the MRZ. PACE, its stronger successor, can also use the card access number (CAN) printed on many ID cards.
In practice the person photographs the document first. The MRZ or CAN is read from that image, and the phone uses it to open the chip. Nobody has to type anything.
Once the chip is open, the phone reads the data groups and the SOD. Passive authentication then checks three things:
The third step is where many implementations get thin. A verifier needs the CSCA certificates of every country whose documents it accepts, from a source it trusts. Master lists, signed collections of CSCA certificates published by national authorities, are the usual way to get them.
If all three checks pass, you know the name, date of birth and portrait on the chip were written by the issuer and have not been changed since. A forger can reproduce the look of a card. They cannot produce a valid signature for altered chip data without the issuing country's private key.
Passive authentication has one gap. It proves the data is genuine, but an exact copy of that data on another chip would pass too.
Active authentication and chip authentication close it. In both, the chip holds a private key that cannot be read out. With active authentication, the reader sends a random challenge and the chip signs it. With chip authentication, the reader and chip agree new session keys, which only works with the right private key. A cloned chip fails either way.
Which protocol a document supports depends on the issuing country and the document generation. Some support both, some only one, some neither. A verification result should say which checks actually ran, so you can tell a clone-checked chip from one that only passed passive authentication.
The difference is the kind of evidence. A photo-only check gives you a probability that a document looks right. A successful chip read gives you a cryptographic statement from the issuing authority that the data is genuine.
It also helps the face match. The portrait in DG2 was captured by the authority, without glare, wear, laminate reflections or the print pattern of the card. Comparing the selfie against the chip portrait avoids most of the noise that comes from photographing a printed photo.
And it catches attacks a visual check can miss: an altered date of birth that was carefully reprinted, a replaced portrait, or a high-quality counterfeit of a document type the template library has seen only a few times.
A chip read proves the document is genuine. It does not prove the person holding it is the holder. A stolen or borrowed passport passes passive authentication perfectly. That is why the chip read is followed by a face match and liveness: the selfie has to come from a real person who is present and match the chip portrait.
There are practical limits too:
A few points that matter in practice:
For Swiss financial intermediaries, the chip read also has a regulatory role. The revised FINMA Circular 2016/7 lets a chip read with authenticity and integrity checks replace the bank transfer on the online route based on an electronic copy of the document (Rz 33.1). This is a summary, not legal advice; check the circular and your SRO's rules for your case.
secuveri opens the chip with keys derived from the MRZ or CAN, runs passive authentication against CSCA master lists, and runs active or chip authentication where the document supports it. The selfie is then matched against the chip portrait, alongside liveness, deepfake and injection checks. On an iPhone, the chip read runs in an App Clip, so the person does not have to install an app. The result shows the outcome of every check that ran.
You can walk through a full onboarding flow in the bank demo, or read more on identity verification.
This article is general information, not legal or compliance advice. Your obligations depend on your sector, licence and markets.
Build a verification experience around your business.

How attackers exploit documents, images and digital identity signals — and where verification stops them.
5 min · October 2026
Read More: The New Anatomy of Identity Fraud
What true accept and false reject rates actually mean when you verify people at scale.
4 min · October 2026
Read More: Face Matching vs. Liveness DetectionTell us what you’re trying to verify. We’ll show you how secuveri can fit your workflow.