This thread is for voicing your opinion and discussing these rules. Please remain civil, though; we know it’s a sensitive topic that has caused quite a lot of bad blood in the past.
There is no perfect way to deal with the issue, but based on our past experience, we believe we have found a good compromise for how to deal with AI usage in the stories we all love to read.
You’re free to voice your unhappiness with the rules, of course, but please understand that at this time we’re not going to make any further significant changes to this approach. If, after a while, it turns out the rules are seriously flawed in any way, we might change them again, but certainly not anytime soon.
In my personal opinion, the label “AI prose” is a bit too much of a blanket statement. I preferred the previous terms because sometimes I would give AI assisted stories a chance knowing people at least put some effort into editing and/or writing portions. Now I’d probably just block that tag altogether, unfortunately. That’s just me, though! I know there have been problems in the past with authors being accused of using AI when they didn’t, so I get that this is in part to prevent things like that from happening.
I’m with Subdominant, I liked ‘ai-assisted’ and ‘ai-written’ because I didn’t mind “hey so I wrote this but used AI as a sort of editor’. It being lumped together as a ‘AI prose’ feels limiting.
Wouldn’t the easiest solution be to simply ban all AI stories? I can’t imagine many readers would be interested in reading something the author couldn’t be bothered to write.
I’d prefer that, but from the original messaging it seems like Corin doesn’t want the site to be a place that’s making a bigger general statement like that. Also, unfortunately from what I’ve seen as long as a story is hot, some people don’t care regardless. That’s why I think having two tags would be best to appease everyone, both people who care about the integrity of the stories posted and who just want to read hot stories.
Where are you drawing the line though buddy ? Even word has AI inbuilt into it ( Co pilot ) if someone was checking spelling , grammar, and then clarity , it might say , this sentence doesn’t make sense here is a choice of 3 sentences (which are AI generated) . Not really thinking, then it could be said their story is now AI assisted.
This seems like a choice to overall veil ai usage which i find disconcerting. With over half of all digital articles online now creating to be ai written i think this more vague approach is only meant to conceal that happening to this site rather than actually interact with it. I think the multiple labels were a solid approach, but this new one is just genuinely one that would move me off this site. I also do think it’s a little telling that while you said readers can complain if they “don’t like something” but are additionally told not to mention something being ai in comments.
So I dont know how real of a question this is but ai checkers and readers looking for ai aren’t going to catch that you previously misspelled words or had small grammar fragment issues. As the post talks about generally this is seen more in the prose. To be fair I don’t think full banning works either but that is because I think thay would create more strain on the readers team than would be reasonable
Obviously the line would be drawn at “generative AI”. GenAI firms have intentionally and successfully muddied the waters to make us forget about the distinction between GenAI and all the innocuous things we used to call “AI” before the 2020s. But spell check/grammar check is clearly something that’s been around for decades, and “here is a choice of 3 sentences” is clearly GenAI.
The trouble, of course, is in detecting the GenAI use…
This whole conversation boils back down to the solution that was developed before to use the separate tags for this type of content (assisted vs mostly/fully generated), in my opinion. I felt it was at least an amicable solution given that AI stories were never going to be banned outright, it seems. As I said before I feel “AI prose” is just way too general given how many distinctions there could be and it muddies the water a lot
My perspective as a former approver trained to detect AI stories may help clear things up here: when we used “AI assisted,” it was always to indicate a substantial amount of generative AI prose. That is, stories which were assisted by AI in undetectable ways, like spellcheck, help with idea development and outlining, etc. were not covered under this tag when we assigned it.
As far as I can tell from these rules, the types of stories that would require “AI assisted” under the previous rules still require “AI prose” under the new rules; no stories that would not be tagged by us as AI will be tagged under this ruleset.
That’s my point though . Generative AI is now embedded in a tool like Microsoft word that was previously a just word processor. So to suggest that any form of AI should be banned . How are you putting the genie back in that bottle
Again, that’s why on this site AI use in stories isn’t banned outright. I’d say the main discussion is about tags for AI stories instead of the use of AI itself since that’s been discussed many times before. As Soren said too, the tags (even AI assisted) were for when stories found to be using mostly AI generated prose were used. The difference now is that all stories (whether it’s a story an author fine tuned after generating with AI or one that was almost entirely generated with AI) are now under one umbrella term.
Yeah , I can’t see the issue going away or getting any less thorny anytime soon. I probably shouldn’t comment on the subject , as I guess I take it personally when people suggest AI = a lack of effort☺️
(Not sure if you noticed, but I’m not the person who said “ban all AI stories” – that was Manicorn. I just replied subsequently to that.)
I mean, as I said in the last line, the genie that can’t be put back into the bottle is not GenAI in Microsoft Word, it’s the problem of detecting the use of GenAI.
As someone who’s anti-GenAI and blocked all the previous tags and is happy to be blocking the new tag, I would love for people using even as little as a single “Microsoft Word suggested generated sentence” to proactively self-tag as AI. But I recognize that’s not going to happen.
I do, however, wish the moderators were willing to take a harder tack and enforce the tag nonconsensually when stories are obviously AI-assisted but the authors have an ego about it, like Wrestlr recently (whom I was so disappointed in! Even more than the AI usage, his tantrum about it being detected was what has made me view his new works on other sites with disgust and a skip rather than excitement and a read, even if they are probably not GenAI like the two that tripped the alarm).
There’s no perfect solution, since GenAI is trained on human writing and thus all of its “tells” are just heightened and extreme versions of what real humans have already been doing. But I do wish that when those tells are loud enough for the moderators to be suspicious and approach the author, that would be publicly visible so I know to skip the story. I don’t want to read those cliches regardless of whether they’re written by a human or generated by an algorithm. The old tagging system wasn’t perfect, but it was a lot better for that than this opt-in-only system.
apologises buddy , it wasn’t meant to suggest that you said they should be banned, it was more a reference to a general point of view that had been put across in various comments .
I feel like a lot of people are really getting into the weeds about this when the actual policy seems pretty straightforward… the most obvious slop (unedited prompt responses) is directly banned and the rest is at the discretion of the approval team, with the hope that the people submitting will be honest. Until AI detection gets more reliable, there’s no realistic way to ask for more than this (plus I don’t really think it’s fair to ask for more anyways since any additional burden would fall onto the volunteer approvers). If something slips through the cracks and you realize the story you’re reading might have AI prose in it, just read something else.
Basically, a story which would be truly “AI created” is no longer allowed on the site at all.
All stories which are “AI assisted” in a way that it’s still not too obtrusive and still results in a pleasurable, creative story, are allowed under the new rules with either the “AI prose” tag or a disclaimer.
