Biometric passports follow ICAO Doc 9303, the international standard that makes a chip issued in Lisbon readable by a border terminal in Tokyo. The chip is a contactless smartcard in the ISO 14443 family, the same radio technology as contactless bank cards, with a designed read range of a few centimeters. What sits on it, and who can get at it, is more precisely engineered than most people assume, and less scary than the sleeve marketing suggests.
What Is on the Chip
The data is organized into numbered data groups. DG1 holds the machine readable zone, the same two lines of text printed at the bottom of your photo page: name, document number, nationality, date of birth, expiry. DG2 holds your facial photo as an actual image file, higher quality than the printed one, intended for automated face matching at gates. Some countries add more: EU passports store fingerprints in DG3, behind extra access control that ordinary readers cannot pass.
Alongside the data groups sits the Document Security Object, a structure containing a hash of every data group, signed by the issuing country. The signature chains to a national root, the Country Signing Certificate Authority, and countries exchange these certificates so each can verify the others' documents. It is a public key infrastructure running quietly across every international airport on earth.
Passive Authentication: Why Forging the Data Fails
When a border terminal reads the chip, it recomputes the hashes and checks the signature. This is passive authentication, and it means the chip's contents cannot be altered without invalidating the country's signature. Change one byte of the photo and the hash no longer matches; forge a new Document Security Object and you need the issuing country's private signing key.
Note the precise claim. Passive authentication proves the data is authentic and unmodified. It does not prove the chip is the original. A bit-for-bit copy of the data onto a blank chip carries valid signatures, which is exactly what security researcher Lukas Grunwald demonstrated in 2006 by cloning a passport chip's contents. The clone defeats nothing on its own, since the copied data still shows the original holder's face, but it exposed the gap: signatures authenticate data, not hardware.
Closing that gap is the job of active authentication and its successor, chip authentication. The chip holds a private key that never leaves the secure element, and proves possession by responding to a challenge. A cloned chip has the public data but not the key, so it fails the challenge. This is the same design logic as a FIDO2 security key: the secret stays in hardware, and presence is proven by signature rather than by data.
BAC: The Weak Lock on the Front Door
Radio interfaces invite eavesdropping, so the chip refuses to talk until the reader proves it has seen the physical document. The first-generation mechanism, Basic Access Control, derives an encryption key from three fields of the machine readable zone: document number, date of birth, and expiry date. A reader that knows those fields, normally by optically scanning the open passport, can start an encrypted session.
The flaw is entropy. Those three fields are structured, not random. Document numbers are often sequential, birthdates cluster over a human lifespan, and expiry dates fall within a ten-year window. Researchers showed in the mid-2000s that the effective key space for some countries fell to roughly 35 bits or less, weak enough that a recorded chip-to-reader session could be cracked offline in hours. Adam Laurie demonstrated reading a UK passport chip in its sealed delivery envelope by exploiting the predictability of the key fields. Where document numbering is predictable, low entropy quietly becomes the whole story of the system's security.
PACE: Fixing It With Better Math
The repair, Password Authenticated Connection Establishment, has been mandatory in EU passports since the end of 2014 and is now the standard access mechanism worldwide. PACE uses the same low-entropy inputs but runs them through a password-authenticated Diffie-Hellman exchange. The session keys it produces are full strength and independent of the password's weakness, and an eavesdropper who records a session gets nothing to attack offline. A wrong guess at the password fails online, one attempt at a time, against a chip that can throttle.
This is the same conceptual move that modern login protocols made: stop treating a weak secret as key material, and use it only to authenticate a strong exchange. The passport system arrived at it a decade before most websites did.
| Mechanism | What it proves | What it cannot prevent |
|---|---|---|
| Passive Authentication | Data is signed by the issuing country and unaltered | Cloning the data to another chip |
| BAC (legacy) | Reader has seen the printed MRZ | Offline cracking of recorded sessions |
| PACE | Reader knows the MRZ or card access number, with strong session keys | A reader that legitimately has your MRZ |
| Active / Chip Authentication | Chip holds the original private key (not a clone) | Physical counterfeiting of the booklet itself |
| Terminal Authentication | Reader is authorized (gates fingerprint access) | Misuse by an authorized terminal |
How Real Is Skimming?
The image that sells shielded sleeves is a stranger on a train wirelessly lifting your identity. Against a modern passport that scenario mostly fails on its own: the chip will not release data groups without the access keys derived from your MRZ, so a skimmer who has never seen your photo page has nothing to say to it. Contactless skimming is a real category, but the passport's access control was designed against precisely this attack.
The residual concerns are narrower. Early chips answered with fixed hardware identifiers before any authentication, which allowed a specific passport to be recognized on reappearance; randomized identifiers have since become standard practice. A device can still notice that some passport chip is present nearby. And anyone who has legitimately handled your open passport, a hotel desk, a landlord, a border agent, has the MRZ and therefore chip access. That last one deserves more of your attention than strangers with antennas: a photocopy of your photo page is the key to the chip, so treat photocopies with the caution the sleeve marketing spends on radio.
Many US passports since 2007 ship with shielding in the cover, and a closed passport with such shielding is already hard to energize. A faraday sleeve defends against a narrow, mostly mitigated threat. The higher-value habits are boring ones: control physical access, be deliberate about who photocopies the photo page, and treat the document number as identifying data, because it is.
Passports also travel with you across borders, where the chip is the least of the exposure. If your threat model includes device searches at entry points, the passport chip is the well-defended part of your kit; the phone next to it is not. That problem has its own playbook, covered in our guide to border crossings and device privacy, and the emerging generation of digital identity wallets is about to reopen every one of these design questions on new hardware.