Topic: RIP Kling AI? (NSFW detection update)

Posted under General

This topic has been locked.

I'm sure people have noticed that Kling's NSFW detection system has gotten an update where you can't upload absolutely any type of NSFW content. I recently started animating my images through Kling, and I was very satisfied with the results most the time, but unfortunately the software got said update before I got to animate all the images I had in mind. The detection system had always been present, meaning that sometimes I had to find some -albeit simple- workarounds to bypass it, but now it seems like the detection system has gotten absolutely op. Now, not even extreme image editing can go through unrestricted.

To be clear; I fully understand the dangers of NSFW AI media of real life people and humans in general, so it does make sense to me why they would make the restrictions stricter. However, when it comes to furry content I feel like the restrictions are quite unnecessary since they are all fictional characters. And let's be real, the Kling team cannot deny that a huge population of their users uses the software solely for NSFW purposes. A huge population that might no longer use/support this software/company. I for one can no longer use it. So they must be losing quite a few customers.

I would love to hear your thoughts on all this, since I can't find any coverage of this online. Especially from the video creating people here. How do you deal with this update? Any thoughts, hopes for the future, and preferably any alternatives to Kling are very welcomed. xx

Updated by Jelloponies

butterlait said:
I'm sure people have noticed that Kling's NSFW detection system has gotten an update where you can't upload absolutely any type of NSFW content. I recently started animating my images through Kling, and I was very satisfied with the results most the time, but unfortunately the software got said update before I got to animate all the images I had in mind. The detection system had always been present, meaning that sometimes I had to find some -albeit simple- workarounds to bypass it, but now it seems like the detection system has gotten absolutely op. Now, not even extreme image editing can go through unrestricted.

To be clear; I fully understand the dangers of NSFW AI media of real life people and humans in general, so it does make sense to me why they would make the restrictions stricter. However, when it comes to furry content I feel like the restrictions are quite unnecessary since they are all fictional characters. And let's be real, the Kling team cannot deny that a huge population of their users uses the software solely for NSFW purposes. A huge population that might no longer use/support this software/company. I for one can no longer use it. So they must be losing quite a few customers.

I would love to hear your thoughts on all this, since I can't find any coverage of this online. Especially from the video creating people here. How do you deal with this update? Any thoughts, hopes for the future, and preferably any alternatives to Kling are very welcomed. xx

There seems to be ways to go around it still, though you'd have to look for them, since this one came out: https://e6ai.net/posts/75534?q=epic_%28director%29 though I imagine they will continue updating their thing to make it harder and harder. It's their choice, their platform, so I assume even firguring out a way to go around this current update will be a temporary solution.
The same way options like Frosting came out for NSFW/furry image generation, I'm affraid it's probably better to wait for a good option made for our kind of content in the end...?

I'm using Soulgen to do the same thing. I believe they must have made some fine-tunings based on Kling or wan2.1. Judging from the results, it seems to be even more outstanding. I'm not sure if it can always perform well for the furry category. Currently, I have a probability of about two-thirds to get relatively good results. They initially offered a free trial, but recently they canceled it, citing the damn costs and the need for faster processing speed. It's really ridiculous.

scalesalad said:
There seems to be ways to go around it still, though you'd have to look for them, since this one came out: https://e6ai.net/posts/75534?q=epic_%28director%29 though I imagine they will continue updating their thing to make it harder and harder. It's their choice, their platform, so I assume even figuring out a way to go around this current update will be a temporary solution.
The same way options like Frosting came out for NSFW/furry image generation, I'm affraid it's probably better to wait for a good option made for our kind of content in the end...?

Nah the site is dead. That was uploaded 5 months ago. They've been steadily ramping up censorship for months and last night they updated once again and have the AI detecting any sexual content or poses, any phallic shapes, orifice shapes, and block any generation at all even if it lets you upload the image and waste your points trying to generate, absolutely nothing is getting through anymore. I have been a subscriber for months now and have just canceled my membership. Its astonishing that the company is this stupid, although I also have to wonder how many idiots are generating porn and then trying to publish it, so some of the blame must go to these people as well.

pohan said:
[...] although I also have to wonder how many idiots are generating porn and then trying to publish it, so some of the blame must go to these people as well.

I don’t like to be called an “Idiot” for sharing some of my NSFW Kling gens with the community.
And also it doesn’t really matter if you publish them or not. Do you really believe the Kling devs are monitoring e6ai to find out if somebody uses their service to gen porn?
No, they don’t have to do that. They can look at everything you upload and generate on their site, whenever you publish it or not (of course they can’t look at everything because it so much).
The easiest way to do this is to just take a closer look at the accounts that have many refusals (yours & mine for example).
And then take the input images we uploaded and finetune their vision-LLM they use for censoring based on our inputs – as examples for what is “unsafe”.

penisfire said:
I commented right here, it got edited out by a mod... seriously stop makinig me hate this place.

I took a look. Nobody has edited your forum post except you.

penisfire said:
Didn't ask for a log post. At the time it said "edited by Angry Puppy". Didn't notice a change, but it was weird to me that it said that.

Absolutely false, I did not edit your message whatsoever.

angry_puppy said:
Absolutely false, I did not edit your message whatsoever.

I saw the name of a mod - edited by Angry Puppy. If not that name then it could be confused with another similar, but 1000% saw it.