This isn't a good idea. It's quite common to use some feature of the underlying hardware when generating a UUID on a platform - MAC address for example - but you have no guarantees.
I've used device ID (a UUID) from iPAQs in some mobile data collection applications to trace where data was collected/uploaded from, and whilst it was usual to find related batches of iPAQs would have similar patterns this would change radically with different models and sometimes within the same model range for no apparent reason.
It's possible to overlook how mind-bogglingly vast the space of UUIDs is. To quote wikipedia
In other words, only after generating 1 billion UUIDs every second for the next 100 years, the probability of creating just one duplicate would be about 50%. The probability of one duplicate would be about 50% if every person on earth owns 600 million UUIDs.
Having said which I did hear that Second Life once encountered a problem with duplicate UUIDs being generated (where every object in the world is identified by a UUID) because of a very very obscure problem in the Python library they were using to generate them, so the risk of collision is probably a mite higher than the theoretical due to coding assumptions.