AAP 2026 CONFERENCE

There are multiple vectors along which one can be more or less pessimistic about the prospects for human-AI friendship. I could be optimistic about the purely technological aspects of AI friends, believing that they will soon be able to do some non-trivial percentage of the things the AI companies claim they can already do. I could be optimistic about the affordances of these AI friends for human relationships and experiences. I could be optimistic about the implementation, or the regulation, of such AI. Here, I examine and taxonomize these possibilities.