This will become a big issue in the future. We can't just recreate the intricacies of a brain and claim it can't be sentient like other comments in this thread.

Yes, at its core it is only predicting the next word it wants to say based on a complex series of weights, but is there any evidence that you are not doing that yourself?