If I understand what you are asking, I think the simple answer would be that AI is nowhere near having demonstrated sentience, thus they do not qualify for any type of rights.
We won't have to "cross this bridge" until an AI demonstrates self-awareness and human-level-or-beyond intelligence, but it sure is interesting to think about!
(Also, the UDHR dates to the 1940's and seems to have had its last additions in 1966. Computers weren't very "smart" back then so likely no on was even considering the question ;)
Although you may also want to look at the grey good scenario, which posits inadvertent destruction of homo sapiens sapiens not as a factor not of too much, but of too little, intelligence.
The problem with an Asimov approach is highlighted by his book I, Robot, which is the potential pitfalls of pure logic. The philosophy of Neo-Luddism is preoccupied with these problems in relation to technology--specifically that the threats posed by technology cannot be predicted.
The problem with the UDHR today is that there is no algorithm smart enough to understand it--we're not even close. (There is something called the symbol grounding problem which demonstrates that meaning and understanding in relation to algorithms is still unsolved.)