Home | Notifications | New Note | Local | Federated | Search | Logout

Note Detail

Woozle Hypertwin@woozle@toot.cat (2026-08-12 22:05:19)
:tootcat: So, we've had a credible report that a bot account on our instance -- @newstldr -- is posting information of questionable accuracy, presumably created by LLM.

I'm trying to decide on an appropriate response.

(For clarity: this is not our bot, I don't know who runs it, I don't know what model it is using or how that model is hosted.)

Some of my thoughts:

Any LLM account should have, in its profile: who runs it, what model it's using, and how that model is hosted
If the information's accuracy cannot be vouched for, then the post should be behind a CW saying something like "LLM-generated news summary, not checked for accuracy"
Any LLM posting factual content should be responsive to corrections in comments (I haven't tried this -- partly because I'm not sure which pieces are specifically inaccurate).
Thoughts?

#TootCat #LLM #AIethics
---Reply---
dave@dthompson@toot.cat (2026-08-12 22:28:15)
@woozle I think a simply "no llm bots" policy would be appropriate.