I'm strongly partial to a wearable token. The NFC Ring is one highly attractive option.
- It's inobtrusive enough to wear all, or very nearly all of the time. Contrast cards or similar carried-but-not-worn tokens.
- It can be readily use to tap a sensor for identification purposes. Contrast cards or similar tokens (e.g., USB keys), which are far less immediate.
- It is replaceable. That is, if it's compromised, stolen, or lost, it can be replaced. If it becomes unadvisable to possess, it's readily discarded and reasonably easily destroyed. This contrasts with biometrics or permanently embedded sensors.
- Its absence is reasonably immediately determinable. Again, contrast carried-but-not-worn tokens.
- The existing prevalence of ring-wearing makes use of an NFC ring less obvious or evident (mostly a concern in early-adoption periods), or the opting-out of wearing one (which ring is the NFC ring?), without directly querying each individual, which ... might not work regardless (depending on implementations).
- There are relatively few people who would be entirely unable to use such a device. Ready alternatives for most such cases exist: wrist bands
- Unintentional validation (e.g., surveillance) is relatively easily avoided, if devices require immediate contact with a sensor/receiver. That is, a surveillance entity couldn't mass scan a crowd or region quickly, but would have to individually query rings in close proximity. (This might be achieved through high-volume transit points already, but this already raises the ante.)
- It's possible with a query/response system that multiple identities with the same root, but not immediately correlated, could be supported. (Deanonymisation or identity linking remains a significant problem, however.) Ideally, such a system could be limited to only satisfying minimum qualifying criteria (e.g., "I've paid a fare for this trip"), rather than transmitting either a full personal dossier or an absolute identity.
Key (so to speak) challenges are in agreeing on a single standard, ensuring crytpgraphic robustness, and protecting privacy, surveillance, and other concerns, as well as distributing the detector infrastructure for desired uses.