Based on recent comments this feels like a discussion we should have. So…topic, basically.
I’m not looking to be chief noisemaker on this, but I stand by what I wrote in !privacy and what’s in my post history.
https://lemmy.ml/post/48724623/26190950
Let’s have at; do we want a [AI] and [NOT AI] tag. Why or why not?
Absolutely.
Please do! It’s always my first question when reading new projects.
A mandatory [AI] tag? Sure.
A [NOT AI] tag? No, that’s the default. Why normalise AI bullshit even further?
But mandating [NOT AI] means that people have to go out of their way to declare their work is AI-free. It requires active lying rather than lying by omission—I think there are a non-zero number of people who would be inclined to omit an AI tag but would not want to go as far as explicitly lying about their work being AI-free.
I think [AI] tags would be good. That way a certain subset of members could just drive-by downvote without getting themselves dirty. [NOT AI] seems redudant since we’ve already defined [AI], but again for quick filtering purposes, I see no harm in both.
That way a certain subset of members could just drive-by downvote without getting themselves dirty.
I think tags could be alright but only if this is not allowed, it is unreasonable to ask people to disclose something just so others can shit on them for it.
it is unreasonable to ask people to disclose something just so others can shit on them for it.
I totally dig what your saying. I’m not a downvoter period. In my short time here at Lemmy, AI assisted projects are going to get shit on one way or another. It’s unfortunate, but it is what it is. I think the narrative of this thread is to attempt to make things conducive to all users. I personally do not outright reject AI assisted projects. My main concern is if I spin up a container and it turns out to be a doughnut. AI assisted or no, unless you speak multi code languages fluently, you are taking a risk either way. You are placing your trust in the dev and the few that can read code.
you are taking a risk either way. You are placing your trust in the dev and the few that can read code.
There is definitely a trust issue and a need for ways of conveying and building trust in smaller software projects. I think a much better solution there would be discussions about the code and how it works that aren’t hostile interrogations with foregone conclusions in pursuit of a broader anti-AI agenda. If someone just put a lot of effort into making something the details of that process should be on their mind, it should be possible to make them more accessible to people and convey that there is non-artificial understanding behind the project. Automatic hostility and suspicion makes those kinds of conversations harder and less likely.
Automatic hostility and suspicion makes those kinds of conversations harder and less likely.
You’re preaching to the choir but I will give an amen.
Having both an [AI] and a [not AI] tag allows immediate differentiation between a not AI post and a did not tag post.
But it’s annoying. Non AI should be default, AI has to be marked.
Reasonable.
I would support those tags. Does Lemmy support some equivalent of post flairs that can be filtered?
Good question - and that’s another problem. Not sure - can’t see any in Voyager. Can you see any on your end?
I did a little digging, and apparently there is a plugin for Lemmy for flairs. https://github.com/basedcount/flair/blob/main/README.md. Again, I have no experience standing up a Lemmy instance nor do I know how hard that would be to implement if that option was favored.
I’ve never seen it on the official web app. I suspect that, if they existed, I would’ve seen them used by now.
yes.
always yes.
Just ban AI already
This is a better plan.
Yes please.
Yes, please. I don’t like seeing a “neat handy application” only to find that 95% of it was coded by Claude, the fact of which is either buried, or not even mentioned until you visit the repo and see that it’s the top contributor.
If you see this seal:
You know the project is golden. It uses a no-ai policy specifically.
https://sciactive.com/human-contribution-policy/
We should be pushing projects to adopt this policy (or similar) to have a very clear indicator when something is not slop.
In addition to this, could we potentially make this a badge as well?
https://brainmade.org/That one is good, but it doesn’t have strict licensing restrictions like the SciActive one. Basically you have to follow the policy with the SciActive one, or you lose the license to use the seal.
Also, the SciActive one is specifically for software projects and covers the full software project explicitly.
Not that the Brain Made one is bad, it just fits a different use case (art, poetry, etc).
I absolutely don’t want Meta Tags in every titles. It makes reading the list of posts super annoying.
I also don’t feel the need to know whether there’s some AI commits, but I do want to know if a project is largely vibe coded. I don’t have an objective metric on where this line could be drawn.
I think the status quo is kinda fine. Some commenter will point it out and will get enough upvotes to be visible on first glance. It’s not perfect but good enough for me.
Some commenter will point it out and will get enough upvotes to be visible on first glance. It’s not perfect but good enough for me.
I don’t necessarily disagree here, however there is always the edge case of a certain overly-vocal group being 100% anti-AI, and that any usage of AI is considered a crime against humanity.
I’m not part of that group per se, but I also want full disclosure. If you used AI to get things going and handled it yourself from there, that’s one thing. Constant commits from Claude or whatever is a whole different bucket of shit I refuse to touch.
I don’t disagree. Perhaps tags at the end of post titles rather than the beginning, or tags in the body of the post? I’m not sure what people use to filter these things and how those different options would effect usability.
I feel like this could lead to discrimination and prejudice against ai users. It should have been implemented years ago.
To clarify…you in favour of discrimination and prejudice against ai users?
Yes. They should be told to sit in the corner with a big cone shaped hat on like the old days.
Yes. Anything made, posted, modified, or output by AI should be tagged as such, always, without exception.
Including immich, home assistant, syncthing, docker…
Right?
I know nothing about running a Lemmy instance, so I thought I’d ask. Does the Lemmy framework allow something like this:
spoiler
If so, maybe that’d be less annoying to some. The screenshot is from selfh.st.
I think that unless you have some way to enforce accuracy, it’s meaningless and AFAIK automatic detection tools are no better than chance and to my knowledge, getting worse.
An AI bot operator isn’t going to tag their material as [AI], more likely than not they’d attempt to use [NOT AI].
I’d also point out that while lemmy doesn’t (yet) support hashtags, any “tagging” would probably benefit from using the existing method using a #tag.
Ultimately, you need to ask yourself, is undeclared AI that goes undetected by the community a problem, or the new “normal”?
I’ll note that I’m not a proponent of Assumed Intelligence and think that when the bubble bursts we’re going to be in a world of hurt, but with a little luck the billionaires will have lost their shirts in the process.
AFAIK automatic detection tools are no better than chance and to my knowledge, getting worse.
It depends.
There are a variety of indicators, some blatant like .cursorrules or CLAUDE.md and some key phrases that can be searched for. Unless they are actively working to hide references, these turn up in the overwhelming majority of the “not admitting but not denying” camp, or the just complete lack of disclosure.
There are also pretty solid indicators with commit logs that can be seen that are… unlikely to ever be a person.
What is getting harder is around the syntax detection tools and fragment detection tools. Some straight up rips from multiple codebases used to be obvious, but some tools are better at refactoring to make that discovery harder too.
Just to be candid, I am in no way anti-AI, I run and train my own models at home on my own hardware. Its a tool - a hammer is great for nailing a board to a wall, and an absolutely wrong choice for trying to screw together a cabinet, and I consider LLMs in the same camp.
If you’re trying to vibe code a full project, its probably going to be a massive problem. If you’re using it to parse some swagger and generate some hooks based on a prefix you defined for specific types without wanting to bother with smashing a script for the disparate types that can be easily detected with an llm… well, you’re probably using it in a way I’d agree with.
Someone did mention an automod disclosure comment, which I really like the idea of, but would need to look into for use on the fediverse, I’m really not sure how good the automods out there are these days. Last time I checked they were… not so great.
