I've been thinking about this without having tried it. Part of me is drawn to the idea of it in a tickling context. It could be used for so many applications related to it. However, another part of me is very uncomfortable with increasing human-to-system interactions and decreasing human-to-human interactions, especially for those interactions that are deeply social and paramount to how we relate to one another on small and large scales. Also, for me personally, there is something about knowing that it's inorganic and inauthentic (i.e. it's not human, but essentially pretends to be) that ruins certain otherwise potential applications for me, regardless of how convincing its presentation is. It's a weird thing to try and get my mind around and to explain exactly how I feel about it. Maybe I'll feel differently in the future, or at least have a better understanding of my own gut feeling about it.
Where I think I would possibly use such technology, though not prefer it, is in converting textually written stories into audio stories. I think in the scope of tickling, that would add interesting color to some stories when it is desired. Again, I'd prefer human readers, but there is cost, rarity, and other challenges to overcome there.
Where I think I would not use it is as a supplemental individual to talk about tickling with, or to role-play with.
I'm really curious what others say to this, as I'm still processing exactly how I feel about AI in many, many contexts.