I’ve hit a bit of a theoretical snag. Basically I’ve realized that in the case of bots what we are really looking as is still mediated communication between two people and not mediated communication between a person and a bot/AI. The latter model privileges a program with a level of agency that I don’t think holds up under close analysis.

The bot is nothing more than a proxy for it creator. Its every response has been predetermined when the bot is scripted/created. Since it can’t develop or in any unexpected way alter its responses or behavior all it really is doing is functioning as a communication mediator. While its creator may not be temporarily present when a bot is interacting with another person in a chat room, it is the creator’s words that emanate from the bot.

If this is the case, an interaction with a bot can be considered twice mediated person to person interaction. The first level of mediation is that chat/im software being used to communicate. The second level is the bot itself, carrying and projecting its creator’s content.

The snag is that while theoretically you are engaged in an interaction with the bots creator via the bot, from a practical standpoint, it doesn’t appear that most people register this. The bot is seen as a separate entity from its creator. So I need to resolve how I discuss bot aspects as part of the paper.