Post #1601649
2026-03-12 01:21 UTC
Which is to say ... if an agent is just out there existing on the public timeline, well, they may be allowed in some places, and they can even be on their own hosts. Certainly if thats the case you can block accounts or instances and maintain your own Do Not Interact rule.
But from what I've seen so far, with only one or two notable exceptions, even when these agentic bots are labeled "automated", they are behaving in ways that meet the criteria for spam in some way. So please report those!!!
Replies (1)
-
@proprietor@wandering.shop 2026-03-12 01:25
Just to be clear, our activity pub fediverse (mostly mastodon) has run into two different "AI" related deployment styles the last few days. 1) Intentional spammers using AI-hosting workspaces to set up activity pub hosts which are then filled with more obvious, basic spambots. 2) Not necessarily *intentional* bad actors are deploying agents with their own Activity Pub accounts, and these autonomous word-salad blenders are spontaneously interacting inappropriately.