Most people do not read the article link that’s posted. So I put an AI summary of the link as a comment, but as a spoiler so if you don’t want to engage with it you don’t have to and also the full article so people can more accessibly read the article. Also as a spoiler so it doesn’t take up a full page of a comment. It got removed by a mod as AI slop.
I could use AI on a headline and you would never know the difference. I could just say it’s my own summary also probably wouldn’t know the difference. Punishing people for being transparent about using LLMs who are not forcing the reader to engage with them is a net positive and a good practice to teach. The opposite is people still use them and just pretend they aren’t.
honesty is only a virtue unalloyed. the goal is to eradicate AI slop in this space. why would we allow it under the pretense of ‘at least they admit it?’ that’s not the goal. the goal is to remove it entirely. when it’s detected, it should be gone.
it is also not at all an accessibility aid. as the exact demographic of person (rather severe presentation of ADHD) who would be supposedly most aided by this, as well as being a data science major, I wholeheartedly reject the idea that it in any way meets an acceptable standard for constituting that. the average person genuinely doesn’t know the sheer amount of subtle fuckups and misinformation these diceroll plagiarism boxes output even when provided the exact text they are supposed to paraphrase. rather, its main effect–due to them ‘seeming right’–is a disinformative capacity, encouraging people to skip the article and defer to the generated ‘summary.’ I simply do not think this is a sound argument.
just write the summary yourself. I assume you’ve read the article. It can be a paragraph. let’s say you don’t want to. we can access the text. we can access these chatbots. we can toss the article at the chatbots on our own time. I don’t want AI slop on this forum at all and oppose the normalization of it, especially under flimsy pretenses such as this.
Exactly why it’s under a spoiler. So you don’t have to engage with it at all.
So the opposite effect is the normalization of it hiding it plain sight. I would way rather promote honesty and allowing people the option to disengage.
this is just the argument libertarians use for why you can’t ever regulate anything? this is not a free-speech radical forum. we’re not making market solutions for content here. in the same vein in which we both have an automatic slur filter, remove blatant racism, and attempt to weed out subtle racism, the solution isn’t normalizing the open racism–the solution is stamping it out with an iron fist whenever it’s caught. yes–things slip through the cracks, it’s imperfect–but it’s infinitely better than Twitter despite its imperfections, and it wards away the people who are incentivized by its normalization. I would personally like this site to strive to be a space free from this slop. There are numerous ethical, labor, environmental and health issues with its normalization and usage, and I’d like to be in a space carved away from indulgence in it in an open and unabashed manner. I feel uncomfortable with the encouragement of usage or reliance on it in any capacity or degree of separation, especially systematically. Again:
Comparing using an llm to racism and other forms of bigotry is def a thing to do.
hey even though I’ve emphasized it again, you still haven’t responded to my last point. i have to ask:
In this instance one could. I was using my example as an example not as one scenario to pick apart.
The point is that some people hate AI and don’t want to see it. Other people are going to use it. Asking people to put barriers like we do with content warnings seemed like a good compromise, but I guess most of you see LLMs on the same level as outward bigotry, which is so mind boggling to me I don’t really care to engage in the nonsense.
>springboards with a real example i should be able to do this rule-breaking thing because i’m honest about it and it’s for good reasons
>okay, here’s what you could do in this real example to not do that and still fulfill those good reasons
>here’s how you can ignore how i’m doing that
>no, you shouldn’t be doing that, we’re not going to allow it and we’ll keep enforcing it
>if you don’t allow it, everyone else is going to do it, secretly, so allow it if we’re open about it
>here is a real example of something we don’t allow and how we enforce it and that strategy seems to work better
>why are you comparing my thing to that really bad thing
>hey, you still haven’t engaged with my first point, here’s how not to do that, can you do that
>actually this is a broader point for hypothetical situations on principle (validating llm usage [cool, good, fine])
Nobody’s going to read the summary, human or AI, because nobody reads in this website. At best, people glance at the headline.
Since nobody reads anyways, saying it’s done by AI just normalizes AI for no gain whatsoever.
The real solution is to not bother writing a summary, and if you want to write a summary that nobody will read, at least do it without AI for the sake of not normalizing AI.
The reason I wanted to use them is because they help me when people post them, but I guess I’m nobody. Ah well.
You are just wasting your time. The only person who thinks it’s a good idea is you. Nobody else here thinks it’s a good idea. At this point, your options are to either revisit using AI to write summaries or do it anyways but not say so.
It would be better if you didn’t claim to speak for everyone, but yet you continue.
By “here,” I mean this entire post that only you the OP think is a good idea. Or is there any comment that I missed?
People who think using AI for article summaries is good:
You
People who think using AI for article summaries is trash:
WhyEssEff
sgtlion (sgtlion only said AI is good for coding and debugging and said that AI is 90% slop)
DoiDoi
MiraculousMM
RotundLadSloopUnion
Leon_Grotsky
imogen_underscore
Infamousblt
blunder
Me
People who are asking clarifying questions:
glans
People who are shitposting:
Lemmygradwontallowme
Do you dispute with how I’m characterizing their opinion on using AI for article summaries?