They get paid to produce confusion.

They make cab rides more enjoyable with their deep conversations.

Good example! I think philosophy's purpose is to clarify and systematize unexplored intellectual areas. I imagine philosophers today are already using AI as intellectual sparring partners. I suppose if energy were cheap, we could run AIs all day long to pontificate like humans and write philosophical tracts on the issues of the day. When that happens, we will see if they say anything of merit.

Mathematicians will soon be left only to conjecture, with proofs being automated. The issue I see is that AI will devise proofs that are beyond our comprehension, since humans are already taxing each other (cf. Wiles, Mochizuki, Perelman, etc.) Once humans lose grasp of the proof, how will they propose new conjectures?