Dissecting Literotica’s AI Policy

[TheWritingGroup]

If @Laurel spends just 5 minutes on each of those, even discounting the rejects, that's almost 20 hours a day.

Speculation is that many stories are autoapproved or autorejected.

 

[Candy_Kane54]

Perhaps the moderator for this forum, @AH_Mod, can create a fourth subforum specifically for AI discussions?

 

[RejectReality]

Something tells me that corralling people who are already pissed off into a downwind, stagnant swamp all their own is going to piss them off even more.

 

[SkyBubble]

LLMs are the worst kind of echo chamber, because they ask up front what you want them to tell you, and they'll do that regardless of whether it's true or not.

I've been in discussions with LLMs where they told me what I was telling them is wrong. Corrected me. I appreciate that.

 

[SkyBubble]

Speculation is that many stories are autoapproved or autorejected.

They pretty much have to be.

 

[StillStunned]

Perhaps the moderator for this forum, @AH_Mod, can create a fourth subforum specifically for AI discussions?

They can be annoying, but look at it this way. We get these forums to play around in, largely without any supervision or limitation. The forums don't add much value for the site owners, I suspect: they're purely for our amusement. In exchange, Laurel and Manu (may they live forever) have designated us as the go-to font of wisdom for anyone with questions.

It's a small price to pay.

 

[big_cane_sugar]

Our entire global society is wrestling with what AI will mean in various aspects of our lives. Literotica is a smut site, probably not the place to look for leadership on an issue like this. I get that it's frustrating but if we ever begin to sort out the more important stuff, Literotica will probably not be far behind.

 

[MillieDynamite]

This would only happen if sentence and paragraph lengths had little variation and were combined with far too many compound sentences.

Probably because the system, whatever it is, thinks it's too "standard."

Yes, and I appreciate that they are. I don't like the flase flag rejections, but there is always something that indicates some amount of AI influence. It isn't always true, however, whatever trips the detection is always there.

Good question, Millie. They're working hard to keep out AI slop --as they should -- but they publish human slop. Or what they determine is "human," anyway. And the system generates way too many false positives (as well as, I assume, some false negatives.)