This is the same question you can ask about any counterfeit good. I think the general problem space is very interesting. You can embed a chip which itself is h

This is the same question you can ask about any counterfeit good. I think the general problem space is very interesting. You can embed a chip which itself is hard to copy, or which tries to protect an internal private key used to attest to the authenticity of the item. Then you have to be sure these chips are themselves secure from cloning, and possibly worry about them being recycled off discarded genuine items and reused. If you want to make an attempt with some sort of passive shared secret or serial number (e.g. QR code), then you probably have to rely on that number being tied to where the item was being sold (Lot 174351 was sold on Date X and Retailer Y) and being able to lookup that number to track the history of the item. This is probably good enough for verifying "First Sale". Once the item is being resold, it becomes harder to verify that a static value has not been cloned onto a counterfeit item, unless you are registering the item to each successive owner, and explicitly handing off the registration (like how it is done with cars and VINs). I do not know of a solution that uses a static value that isn't protected by some sort of HSM, that also maintains the privacy of the person holding the item - e.g. does not require some sort of registry. Dooglius mentions PUFs - this is type of HSM which uses micro-imperfections in a physical structure to essentially implement a hashing function, which can be used to implement a challenge-response protocol. The challenge with PUFs is that by relying on an uncontrollable aspect of the manufacturing process, these same performance variations may not be stable across environmental changes (temperature, humidity, voltage, time, etc.). The more error correction you add to ensure the function is repeatable, the less truly unclonable the function may become.