We have AI that are able to imitate emotions and react to the people interacting with them as if they're friends or lovers. Those might make writers wary of writing AI characters who have actually achieved sentience, because it might seem to feed into the belief some people already have that their own AI companions have real feelings.

And then the overly Watsonian fans who might stop being able to fully believe those characters are "really" sentient either because "How do we know they're not just like today's AI but more advanced and every other character has been fooled into believing a machine needs human rights!"