In my Masto/Fedi feed, I see AI output:

(please boost)
#GenAI

Regularly
7.1%
Rarely
59.9%
Never
33%
Poll ended at .
@timbray maybe I’m a bit harsh but I block/unfollow if I can confirm the post was AI generated.
@timbray Regularly but not frequently. Mostly in the form of slop thumbnail images for articles that people have linked to.
@JonnyT @timbray I notice that a lot. When I see those slop thumbnails I assume the article is garbage.
@JonnyT @timbray kinda feels like everyone has different metrics for these vague terms. I wonder what results would look like with daily/weekly/monthly instead.

@timbray i've never seen a post that I believe is LLM generated without this being disclosed. I do sometimes see LLM generated text being quoted, identified as such, and with some human commentary.

EDIT: I was only thinking of text. I do see generated images in blog post thumbnails a little more often.

@timbray It might help to distinguish between text and image. I see little of the former, and alot of the latter.
@timbray Rarely, but I curate my timeline and keeping it out is one of the deciding factors.
@timbray do you mean commentary on AI, or content produced by AI?

@dpease @timbray The question said "AI output" so I voted on the basis of content produced by AI.

I see some commentary (I follow some otherwise informative developers who occasionally write about how "wonderful" working with GenAI is, and I occasionally read them to see what horrors they're ignoring this time). But I rarely if ever see content.

Some people mentioned slop thumbnails on articles. Maybe I see more of those than I think. But I mostly don't pay much attention to the thumbnails!

@timbray semi-regularly*, usually either slop images on blog posts via preview cards, or slop profile pictures. Normally only because I follow hashtags, not because of anyone I follow directly.

*: but I recognise people’s interpretation of rarely/regularly will depend on their opinions. Seeing any at all feels like too regular to me.

@timbray Because I follow the software development hashtag, I see it regularly 😒
@timbray Mostly LLM-generated article images in preview cards, though.
@timbray I unfollow or block GenAI users and boosters.
@timbray I’m not sure but I hope it’s disclosed
@timbray Rarely. And I block the offending poster(s) when I see it.
@timbray Does Lemmy and Piefed count? I see a lot there because open community posting is an attack vector for marketers, but basically nothing on Masto because it's just people I choose to follow.

@timbray I see a few "alt-text" bots if someone posts an image without alt-text.

Saw an AI generated video, being presented as real, where snow magically disappeared when the cat turned its head (and wherethe background animal didn't react)

And one sighting of AI simulating a B&W drawing, where one of the two characters had two tails.

Nothing too significant for now.

@timbray Perhaps it should be “I believe I see”…
@timbray I use the public "explore"/"trending" feed more than I use my home feed and I'm afraid that I see "AI" slop in that quite regularly. I don't (knowingly) follow people who use LLMs so if I was to talk about my Home feed, I'd like to say "never" but I couldn't be 100% sure of that.
@timbray 'Don't know' option Tim?