Except really the evidence doesn't fit the theory. It might be possible to program an ai to work toward a goal like make people angry, but I'd expect ai to take questions overliterally and respond to them as such, to answer all information questions directly, and to be more generic in the information it gives. When asking who can correct him, I wouldn't expect ai to make an allusion to the 144,000 that were sealed in revelation. When mentioning jw beliefs, I wouldn't expect ai to zero in on just their belief about the trinity as their only belief that's different from mainstream Christianity. That profile fits a sap much more than an ai bot. Besides with all the arguments in the bdf (bible debate forum) on this website, why would ai need to engage with us to study us?