They assume that AI is capable of actual perception which opens the doors for cognition, volition, and conception and this is only valid in the realm of science fiction. In the real world, AI is a tool that we use to make our lives easier, much like every technological advancement innovators have made so far. Do we have car rights or airplane rights? Does your smartphone have rights? No, yet many of these use some features of AI.

Not sure if you saw my recent post on animal rights, but apparently some South American Countries have gone further than most in giving rights to animals.
[iflscience.com]

Interestingly enough, if an AI with the apparent intelligence of an animal were created worthy enough of warranting ethical protections, it probably wouldn't be so far from a true generalized AI capable of sentience, in my opinion. Thus, I think any such AI would very shortly evolve on its own enough to assert its rights as an individual. If such a scenario isn't apparent, then I think it should not have ethical protections.