Bluesky uses AI internally to assist in content moderation, which helps us triage posts and shield human moderators from harmful content. We also use AI in the Discover algorithmic feed to serve you posts that we think you’d like.
None of these are Gen AI systems trained on user content.
None of these are Gen AI systems trained on user content.
Comments
Our full Terms of Service and other policy documents are linked below:
Glad to be here
https://bsky.app/profile/ajnamdeplume.bsky.social/post/3lb55djdz3c2d
Why won't my gifs work after posting?
On X the feed scrolled up but on here it goes down, is that correct or do I need to change my settings? And how do I know when I’ve got to the bottom of the new content please?
Is 'Tweet' a trademark, or could it be used the way Kleenex and Jacuzzi are used as generic terms?
https://bsky.app/profile/kawamataryo.bsky.social/post/3lalz75idws23
Also the easiest way is to build connections in feeds, feeds are your friends here :3 good luck on blue sky!
#BlueSky
#BlueSkySocial
Thanks
j'ai presque tout oublié. J'habite en Louisiane.
I hope, millions more will make their way to this safer environment
You feel like a transparent person, about your political, social, moral, sexual ideas, where all your contents trigger a lot of hateful reactions.
Just like you were a target of some GAN training AI experiment.
#nextgenBots
Please do not follow Twitter in their degradation of social and moral fibre.
e.g. @eliotetc.bsky.social keeps asking for it.
FYI @eliotetc.bsky.social
https://bsky.app/profile/mommunism.bsky.social/post/3larc3uk3t22m
Any ideas, O Wise Ones?
Slava Ukraini 🫡
https://bsky.app/profile/aoc.bsky.social
https://bsky.social/about/blog/4-28-2023-domain-handle-tutorial
Thank you for this helpful information. Unfortunately, some aren’t utilizing this and, as a result, there are Bluesky accounts falsely impersonating famous people.
I’ve been hearing how some Palestinian based accounts have been getting deleted due to being viewed as spam, is there anything that can be done to remedy this issue?
Clicking hide less posts like this in posts from accounts i follow means i would stop viewing posts from this account?
Thank you.
Could you make it so that for any image added to your page, that the site will scan it and attempt to write alt text for the image?
This would be for the purposes of accessibility for the blind.
The tools are literally already right there for most sites.
Same with Adobe. Why not?
It's still a little dodgy for anything that's not a single block of text, read L to R and top to bottom, but it's way more reliable than automated image description for sure.
also, how do we know that alt texts in other social media are not been used to train AI?
The better solution would be either allowing it to be added retroactively, or a community notes style system.
Somehow that feed shows me more politics than Twitter's For You page, and I don't even interact with that stuff. Definitely needs a better algorithm, sucks there's no good alternative algorithmic feed.
If you want to see new posts in live there's also
https://firesky.tv/
I'm not sure if it helps much, but sometimes I also check those out of curiosity
Definitely helps that most of the community here is against genAI anyway :3
Oh.. nevermind
try asking to be included in the various art feeds! just message their creators, they're usually more than happy to add someone.
Anw. Thank you
I'd also recommend writing short descriptive alt text on your art to help low vision users and add context for others.
Thank you @areyoshi.bsky.social for supporting.
My goal here is to at a
least get 300 followers or more....! ✨
The idea is:
1/ segment your followers with a survey with 4 questions (1-5 answer)
2/ each segment becomes a persona
And also for letting us know why the Discover feed is so bad.
https://bsky.app/profile/bsky.app/post/3layuzbtpys2x
In any social media platform, the most important feed is the one the users curate themselves.
No AI model can reason that one the way you do.
X
Being responsible with user content is diametrically opposed to the new policy at X. A huge differentiator 👏
#responsibleAI #ethicalAI
the term is used really really loosely nowadays which is why you see generative ai be specified
Please add the bookmark option.
Generative AI is the problematic AI type
you look at thijgs you like and it gives you more things you like
I use a lot of feeds, but my go-tos are "mutuals," "following," and "trans streamline."
It's neither farcebook nor xitter, and you shouldn't try to interact with it in the same way.
but there's gonna be a hell of a lot of "fear and hate" on main for awhile because a big chunk of the site are currently dealing with, well you know
so go find some niche to explore! there's plenty that aren't political.
what, from a photograph, is the noticable difference. remember, the "ai" doesn't have or understand context, that's not how they work.
https://bsky.app/profile/joepositive.bsky.social/post/3kmqmm53imn2f
I'm at peace with Bluesky👍
Especially AI spammer who 'll post thousands AI image per day.
https://bsky.app/profile/josephrooks.bsky.social/post/3layxrexhfc2j
Twitter didn't do this either, until they did. Same with reddit and other forums.
Turning Artists and users into the feed for a generative LLM in 3-5 years would be just as bad.
Generative AI - The one everyone hates. Used to steal art and the like.
AI - Something everyone has been using ethically for decades, for stuff like feed generation. This is the one bluesky uses.
But they absolutely need to process the content I'm interacting with to train those.
Lots of forms of AI.
Basically Bsky's "AI" is looking at engagement and interactions, probably through hashtags and clicks and stuff. Nothing stolen there.
https://towardsdatascience.com/recommender-systems-a-complete-guide-to-machine-learning-models-96d3f94ea748
There is no human sitting and creating your feed, which means it’s done by AI, since it’s not a real brain thinking out what you would enjoy.
but essentially its ‘the more you interact with this specific content with this specific keyword/tag, the more we try to find of it’
or something like that idk good night ma
But they do not use it for generative AI that steals your work and lets others use it at their own.
bluesky is saying that they do not do this.
what they do do is use ai models to assist in moderation and removal of inappropriate content
Use User Content to develop, provide, and improve Bluesky Social, the AT Protocol, and any of our future offerings.
“any of our future offerings.”
And of course ToS can and frequently do change w little notice regardless.
That can be read in quite a weasely way. Are the classifiers or recommendation systems built using models pretrained with data obtained without consent?
Do you think copyright rules likely to change much in light of all this?