I highly doubt that there are any "pure AI" actors here (assuming currently available technology). I've experimented with LLMs (which are "today's current AI trend"), and they just cannot drop into the conversation and make it look human. There's no engagement, discussion, or emotion—just a series of dry statements, and that would be quite easily spotted. There's also the case of a pre-trained AI model that is trained on mainstream data that is not aligned with the views and values of the community; it'd be funny to see some AI account on a "climate change" crusade here, yielding "scientific consensus" over and over (as it is currently just that: a clever output generator based on some input data).