There is no reason for a bot to be able to access or post on federated social networks if the goal is to make social media humane.
For this reason, bots should be heavily disallowed from posting content to or accessing content from federated social media.
Two examples of bots I have found helpful in the past:
A bot that takes book names and gives the description and author (and maybe a link to goodreads or other)
a bot that takes Lego numbers people posts and gives the name and link
The bot that can set reminders (though I don’t know if it worked)
the one that tells you if your comment is in alphabetical order
The bot that ranked other bots
Wasn’t there an xkcd-ass bot?
The tldr bots were awesome
A bot that masquerades as a regular user and posts paid marketing content once a month
A bot that posts political articles with tailored sensational/biased headlines
A bot that records all activity on an instance and creates a topological graph of the activity tied to users and the content they interact with
A bot that posts a picture of cat, a picture that anyone can fucking google or something that nobody needs or asked for
Who fucking cares if you can come up with harmless “good” uses for bots… they always come with the bad.
Sounds a lot like people.
That third bot sounds really cool honestly. Possible privacy concerns, but not necessarily.
As long as it isn’t the shakespeare bot, it’s negotiable