|
|
|||
| forums: groups: | |||
|
"AI Attitudes Survey 2025"
96 posts 3085 views
Love you, too
CreamMeAgain said: Okay I was staying out of this conversation but seriously this entire thread is a sausage fest of men talking about women's images, not even their own images. Which makes the entire conversation completely tone deaf. Only one other woman has commented on this thread and this conversation needs more women's input and perspective. So here it is.
AI is definitely the new technology that we need to learn to accept as a reality. It's being implemented and it's already replacing workers in the workplace and negatively impacting the job market. People's livelihoods are being affected by this. So there's a human element here that needs to be respected and I think this is what makes AI such a touchy subject.
My biggest issue with AI is the same issue with piracy. It's very easy for people who watch porn to separate the image of a woman from the living human person she actually is. So because piracy is an issue, so is a tool like AI that makes pirating much easier. As far as celebrities, their entire career is actually based on their image and the use of it. Women who would never make porn getting cut and pasted into porn is an age old practice that all women fight against, it violates the rule of consent.
Okay, here's where I get really serious. You're raping a woman of her control over her own body, and the narrative of her own sexual identity and experiences when you steal her image. This is a form of rape, it needs to be called out for what it is. Forcing anyone to participate on any level in your sexual desires against their consent is a form of rape. It's not just an image of a person, it's a real person's image. And it's negatively impacting the person psychologically, which affects a person physically.
It's not okay to use a celebrity's image without consent for public display on a fetish or adult site. It's not okay to use any person's image in this way who doesn't consent to it. And because social media has created an environment specifically for people to display their image, it's really important to respect and protect every person's image as a living entity. It's not just an image, it's a real person.
There is obvious benefit to certain groups of individuals using it, but as a whole, I don't see the benefit of selling AI content. Everyone has access to AI generators. Anyone can create AI. Anyone can create from it and create what they want, they don't need to pay others for it. So I don't see it as a need that only specialized people can provide.
There's nothing wrong with creating a fantasy if AI is used respectfully. But that's the biggest glaring problem with AI, the technology itself isn't as dangerous as the people using it. And that's where I oppose the use of it here on this website. There is an epidemic in this community of piracy, and it's those same people who are looking at AI to make stealing easier for them.
AI is definitely the new technology that we need to learn to accept as a reality. It's being implemented and it's already replacing workers in the workplace and negatively impacting the job market. People's livelihoods are being affected by this. So there's a human element here that needs to be respected and I think this is what makes AI such a touchy subject.
My biggest issue with AI is the same issue with piracy. It's very easy for people who watch porn to separate the image of a woman from the living human person she actually is. So because piracy is an issue, so is a tool like AI that makes pirating much easier. As far as celebrities, their entire career is actually based on their image and the use of it. Women who would never make porn getting cut and pasted into porn is an age old practice that all women fight against, it violates the rule of consent.
Okay, here's where I get really serious. You're raping a woman of her control over her own body, and the narrative of her own sexual identity and experiences when you steal her image. This is a form of rape, it needs to be called out for what it is. Forcing anyone to participate on any level in your sexual desires against their consent is a form of rape. It's not just an image of a person, it's a real person's image. And it's negatively impacting the person psychologically, which affects a person physically.
It's not okay to use a celebrity's image without consent for public display on a fetish or adult site. It's not okay to use any person's image in this way who doesn't consent to it. And because social media has created an environment specifically for people to display their image, it's really important to respect and protect every person's image as a living entity. It's not just an image, it's a real person.
There is obvious benefit to certain groups of individuals using it, but as a whole, I don't see the benefit of selling AI content. Everyone has access to AI generators. Anyone can create AI. Anyone can create from it and create what they want, they don't need to pay others for it. So I don't see it as a need that only specialized people can provide.
There's nothing wrong with creating a fantasy if AI is used respectfully. But that's the biggest glaring problem with AI, the technology itself isn't as dangerous as the people using it. And that's where I oppose the use of it here on this website. There is an epidemic in this community of piracy, and it's those same people who are looking at AI to make stealing easier for them.
I had a reply to this which I decided to create as a separate post since it had less to do with the post and poll:
https://umd.net/forums/wam-models-victims-of-nonconsensual-ai
messg said:
There are 2 areas that I think will have different levels of impact in the near term.
The quality of Deepfaking is getting better and will continue to do so. For imaging, you would previously need to train a local model lora to produce something convincing, it takes time, effort and knowledge to this well but once trained can be applied to any number of images. The quality of this done right is astounding with new models like Qwen, HiDream and Wan2.2 and I've no reason to think this won't continue to get better.
What has happened in the last 6 months is image-edit models have become available that makes it almost trivial for anyone to upload a starting image and ask the model to add mess, change poses, undress etc. Nano banana was the first but qwen-edit is arguably far more powerful from a deepfaking point of view as it is also trainable on any concept or person. By the end of the month, I expect nano banana pro to be released running on gemini 3 pro and this will be be far more powerful than the current model and will support much higher resolution/detail.
This for me is the immediate and most concerning aspect of AI and where we as community need to establish strong guidelines of what is acceptable.
From a general content production point of view and the impact on producers and/or whether one person or another enjoys it.. well I think we've a way to go at least for quality long form video.
I've spent a lot of time training image models over the last year or 2. The quality of trained open source models surpasses commercial models but even if the quality of a single image is indistinguishable from "real material" it is a oneshot affair. There is no temporal progression and consistency, I can't create an image set that is consistently and progressively messy. Edit models may help that as they improve but i have my doubts this is the correct solution.
Video, I disagree that any wam video produced by AI is "good enough" yet. I've mentioned the constraints and limitations of commercial models already elsewhere but they've just not been trained on enough messy content to be good. Grok and VEO3 are quite woeful even outside the guardrail constraints. Sora2 pro is probably the best model for wam content generation but even that struggled with artifacts and inconsistencies even before the guardrails were raised.
I'm not naive enough to say video won't get better but the cost and resources needed are still massive constraints and these models are only producing 8-15sec clips.
Opensource models have the same strengths as open image models in that they are trainable but the resources and skills to train a good model, let alone generate a short 5-8 sec clip without audio is beyond most people. Wan2.5 is so far beyond consumer resources it's laughable and good luck training that on cloud without massive bills.
The nature of technology is that hardware becomes cheaper over time but the models also need to become bigger and better which negates any hardware benefits. There may be a paradigm shift in transformer AI which flattens this out but I would be genuinely surprised to see anything like 1-2min clips in a long time and even then so constrained by guardrails or quality they should have little to no impact on producer Wam content.
johnnypie said: I agree with most everything said. Just realize we're getting into unchartered waters. Up to now we've mainly had to deal with posting one girl's face on another's body; cue in deep fake porn. But to actually produce WAM videos you need to hire models.
Coming shortly, anyone with a modest amount of money (a mere fraction of what it would cost to hire models) will be able to type in their dream script and prompt "turn this into a video" - then sell it to anyone willing to buy.
Coming shortly, anyone with a modest amount of money (a mere fraction of what it would cost to hire models) will be able to type in their dream script and prompt "turn this into a video" - then sell it to anyone willing to buy.
There are 2 areas that I think will have different levels of impact in the near term.
The quality of Deepfaking is getting better and will continue to do so. For imaging, you would previously need to train a local model lora to produce something convincing, it takes time, effort and knowledge to this well but once trained can be applied to any number of images. The quality of this done right is astounding with new models like Qwen, HiDream and Wan2.2 and I've no reason to think this won't continue to get better.
What has happened in the last 6 months is image-edit models have become available that makes it almost trivial for anyone to upload a starting image and ask the model to add mess, change poses, undress etc. Nano banana was the first but qwen-edit is arguably far more powerful from a deepfaking point of view as it is also trainable on any concept or person. By the end of the month, I expect nano banana pro to be released running on gemini 3 pro and this will be be far more powerful than the current model and will support much higher resolution/detail.
This for me is the immediate and most concerning aspect of AI and where we as community need to establish strong guidelines of what is acceptable.
From a general content production point of view and the impact on producers and/or whether one person or another enjoys it.. well I think we've a way to go at least for quality long form video.
I've spent a lot of time training image models over the last year or 2. The quality of trained open source models surpasses commercial models but even if the quality of a single image is indistinguishable from "real material" it is a oneshot affair. There is no temporal progression and consistency, I can't create an image set that is consistently and progressively messy. Edit models may help that as they improve but i have my doubts this is the correct solution.
Video, I disagree that any wam video produced by AI is "good enough" yet. I've mentioned the constraints and limitations of commercial models already elsewhere but they've just not been trained on enough messy content to be good. Grok and VEO3 are quite woeful even outside the guardrail constraints. Sora2 pro is probably the best model for wam content generation but even that struggled with artifacts and inconsistencies even before the guardrails were raised.
I'm not naive enough to say video won't get better but the cost and resources needed are still massive constraints and these models are only producing 8-15sec clips.
Opensource models have the same strengths as open image models in that they are trainable but the resources and skills to train a good model, let alone generate a short 5-8 sec clip without audio is beyond most people. Wan2.5 is so far beyond consumer resources it's laughable and good luck training that on cloud without massive bills.
The nature of technology is that hardware becomes cheaper over time but the models also need to become bigger and better which negates any hardware benefits. There may be a paradigm shift in transformer AI which flattens this out but I would be genuinely surprised to see anything like 1-2min clips in a long time and even then so constrained by guardrails or quality they should have little to no impact on producer Wam content.
I'm also playing around with AI and certainly one thing that will keep the community and producers safe for a while is indeed the guardrails. The more "porn'ish" you try to make the video, the more the guardrails go up. So if you want a "vanilla" pie fight, that's gonna be fine. A topless pie fight? Nope, now you're off to adult AI platforms where it would cost a small fortune to produce videos like that.
PiedVictim said: Truth be told, AI lacks that Human touch.
Sure folk can (and have) always say "its the new normal", "its cheaper and better", "its quicker and easier" or whatever reason to try to validate their justification of using it in any form.
But really, in my view, I see them excuses of not doing the arts yourself. Because anyone can do this (in our case, be it self pie-ing or going out there and interacting with others interested in the same fetish and like pie each other or whatever).
Feels like using AI is a refusal to get better at said thing that is generated by a machine. Want to draw? Then draw. Want to write a story? Then write it. Want to make a video? Then make a video. Want a cute dom woman to smash a pie in your face? Then go out there, make connections/relationships and fulfill that.
It doesn't matter if the works aren't perfect (no one is at the beginning honestly), the point is that there was actual effort to making it, learn your mistakes and that you keep going at it again. With patience, eventually you'll improve and be the ideal artist you wanted to be.
But that's only actually possible if you don't waste your time on using AI Slop.
Sure folk can (and have) always say "its the new normal", "its cheaper and better", "its quicker and easier" or whatever reason to try to validate their justification of using it in any form.
But really, in my view, I see them excuses of not doing the arts yourself. Because anyone can do this (in our case, be it self pie-ing or going out there and interacting with others interested in the same fetish and like pie each other or whatever).
Feels like using AI is a refusal to get better at said thing that is generated by a machine. Want to draw? Then draw. Want to write a story? Then write it. Want to make a video? Then make a video. Want a cute dom woman to smash a pie in your face? Then go out there, make connections/relationships and fulfill that.
It doesn't matter if the works aren't perfect (no one is at the beginning honestly), the point is that there was actual effort to making it, learn your mistakes and that you keep going at it again. With patience, eventually you'll improve and be the ideal artist you wanted to be.
But that's only actually possible if you don't waste your time on using AI Slop.
The thing is, using AI can free up time that you'd rather spend on your other hobbies.
I play D&D, Vampire: The Masquerade, Pathfinder, and some other tabletop RPGs. I run longterm D&D campaigns and oneshots. I write the backstories and lore, plan the sessions, create NPCs and monsters, come up with the encounters my players will face, create the battle maps, and all the other stuff that goes into planning a session. I homebrew (create myself instead of using a pre-written module) 75+% of everything I use in my games. Then I use AI for the tokens, scenery, and NPC pictures since we play online and I can't use the miniatures I've collected or painted myself.
If I had to sit down and create all the art myself, I wouldn't have time to do all the other session prep. And as I've mentioned previously, I don't have money to commission art, especially not for things that will only be seen for a half hour or so and then never seen again.
I'm a very creative person, but my talents lie in writing and imagination, not in creating visual art. There are still some art things I do myself that I won't use AI for (using Inkarnate to create maps, doing hand-drawn sketches for immersive handouts, etc).
The AI WAM images I've shared on here are only a portion of the AI images I've made, and it's all for fun to spend some down time on. Fashion is a passion of mine (not so much keeping up with current trends or designer monstrosities, but with my own style and the styles I like to see), so I spend a lot of time using AI to create the looks I like. Often I'll transition from the style itself to getting it wet or messy, but I usually start with clean, dry images. I also create character concept images for NPCs or characters I might someday play, create dumb pun images, and turn my lame jokes into images.
Calling AI users lazy is disengenious. Yes, it's a stereotype and stereotypes often have a strong base in reality, but a lot of people are like me and use AI for fun or to fill in the gaps, not to do all the work for us. Others, like messg, get really in-depth with the inner workings of the more complicated generators, making learning and practice a hobby in itself.
larryniven said: By "assume," what do you mean? Because what I did was more like "conclude on the basis of evidence." As I said before, I'm going off of the poll results, which are empirical evidence and not, like, some kind of sinister bias of mine. Again, I really need you to read carefully and respond to what I'm actually saying instead of doing whatever this is.
That "empirical evidence" says that 21% of the respondents are willing to buy AI content or pay to create it themselves. That's one out of every five respondents, which is a significant percentage to disregard in favor of the majority. I don't think anyone's advocating for AI to replace real WAM, but there are a lot who are advocating for this site to take a hardline stance against AI.
Kabe22 said:
That "empirical evidence" says that 21% of the respondents are willing to buy AI content or pay to create it themselves. That's one out of every five respondents, which is a significant percentage to disregard in favor of the majority. I don't think anyone's advocating for AI to replace real WAM, but there are a lot who are advocating for this site to take a hardline stance against AI.
larryniven said: By "assume," what do you mean? Because what I did was more like "conclude on the basis of evidence." As I said before, I'm going off of the poll results, which are empirical evidence and not, like, some kind of sinister bias of mine. Again, I really need you to read carefully and respond to what I'm actually saying instead of doing whatever this is.
That "empirical evidence" says that 21% of the respondents are willing to buy AI content or pay to create it themselves. That's one out of every five respondents, which is a significant percentage to disregard in favor of the majority. I don't think anyone's advocating for AI to replace real WAM, but there are a lot who are advocating for this site to take a hardline stance against AI.
So first, to reiterate: nobody needs to specifically try to have AI overrun natural material in order for that to be the result. As I explained earlier, AI-generated material has already proven that it can crowd out natural material simply by virtue of being easier to produce. This gives us a reason to put some protective measures in place even in the absence of, like, a specific intention to have AI replace or out-compete natural material.
Second, whether or not people are advocating for it, people are predicting that AI will radically reshape the economics of WAM production. The second post in this thread, for example, says this:
johnnypie said: When AI get advanced enough, and it will get there, no one will have to pay anyone for anything. They'll create their own WAM dream videos themselves with simple prompts.
Does this count as advocacy? Probably not, but I don't think it matters. What matters is that we're discussing an economic threat to an activity that most of us value. So, again, the economics of the situation give us a reason to put protections in place regardless of whether or not anyone is intending or advocating for AI to replace natural WAM material.
We can talk about how realistic these scenarios are. We can talk about the types of protections that are useful or appropriate. But we cannot, if we are trying to be rational, just say, "Well no one is specifically trying to cause a problem" and then change the subject. That's not a reasonable way to approach this conversation.
As for the question of whether or not UMD should take a "hardline stance," I myself have not seen, like, a huge groundswell of support for banning AI, either from users or from site admins. If you can quote me some people who've said that, that'd be helpful.
larryniven said:
So first, to reiterate: nobody needs to specifically try to have AI overrun natural material in order for that to be the result. As I explained earlier, AI-generated material has already proven that it can crowd out natural material simply by virtue of being easier to produce. This gives us a reason to put some protective measures in place even in the absence of, like, a specific intention to have AI replace or out-compete natural material.
Second, whether or not people are advocating for it, people are predicting that AI will radically reshape the economics of WAM production. The second post in this thread, for example, says this:
Does this count as advocacy? Probably not, but I don't think it matters. What matters is that we're discussing an economic threat to an activity that most of us value. So, again, the economics of the situation give us a reason to put protections in place regardless of whether or not anyone is intending or advocating for AI to replace natural WAM material.
We can talk about how realistic these scenarios are. We can talk about the types of protections that are useful or appropriate. But we cannot, if we are trying to be rational, just say, "Well no one is specifically trying to cause a problem" and then change the subject. That's not a reasonable way to approach this conversation.
As for the question of whether or not UMD should take a "hardline stance," I myself have not seen, like, a huge groundswell of support for banning AI, either from users or from site admins. If you can quote me some people who've said that, that'd be helpful.
Kabe22 said:
That "empirical evidence" says that 21% of the respondents are willing to buy AI content or pay to create it themselves. That's one out of every five respondents, which is a significant percentage to disregard in favor of the majority. I don't think anyone's advocating for AI to replace real WAM, but there are a lot who are advocating for this site to take a hardline stance against AI.
larryniven said: By "assume," what do you mean? Because what I did was more like "conclude on the basis of evidence." As I said before, I'm going off of the poll results, which are empirical evidence and not, like, some kind of sinister bias of mine. Again, I really need you to read carefully and respond to what I'm actually saying instead of doing whatever this is.
That "empirical evidence" says that 21% of the respondents are willing to buy AI content or pay to create it themselves. That's one out of every five respondents, which is a significant percentage to disregard in favor of the majority. I don't think anyone's advocating for AI to replace real WAM, but there are a lot who are advocating for this site to take a hardline stance against AI.
So first, to reiterate: nobody needs to specifically try to have AI overrun natural material in order for that to be the result. As I explained earlier, AI-generated material has already proven that it can crowd out natural material simply by virtue of being easier to produce. This gives us a reason to put some protective measures in place even in the absence of, like, a specific intention to have AI replace or out-compete natural material.
Second, whether or not people are advocating for it, people are predicting that AI will radically reshape the economics of WAM production. The second post in this thread, for example, says this:
johnnypie said: When AI get advanced enough, and it will get there, no one will have to pay anyone for anything. They'll create their own WAM dream videos themselves with simple prompts.
Does this count as advocacy? Probably not, but I don't think it matters. What matters is that we're discussing an economic threat to an activity that most of us value. So, again, the economics of the situation give us a reason to put protections in place regardless of whether or not anyone is intending or advocating for AI to replace natural WAM material.
We can talk about how realistic these scenarios are. We can talk about the types of protections that are useful or appropriate. But we cannot, if we are trying to be rational, just say, "Well no one is specifically trying to cause a problem" and then change the subject. That's not a reasonable way to approach this conversation.
As for the question of whether or not UMD should take a "hardline stance," I myself have not seen, like, a huge groundswell of support for banning AI, either from users or from site admins. If you can quote me some people who've said that, that'd be helpful.
I'm not suggesting changing the subject. I'm actually in favor of:
Requiring all AI content to be tagged as such (with the exception AI used in the editing of real videos/photosets, within reason) The AI filter, so those who don't want to see it don't have to AI content in the forums being limited to the AI WAM group or maybe a more dedicated subforum if AI gains more traction on this site. A complete ban on AI images/video/etc of real people without the subject's consent A complete ban of copyright-infringing AI A ban on links to off-site media that violate either of those other bans. As someone who posts exclusively AI images/gifs, I have yet to have site admin/mods block/deny anything I've posted, so I'm not concerned about them trying to ban AI.
After re-reading both this thread and the one about Robby/Photoslop, I didn't find as many supporting a general ban as I remembered (most were talking about banning non-consentual AI, which I agree with). There were these, though:
PotatomanK said: AI is going to be a problem for years to come. It's entirely trained off of other images. You could put in general prompts and end up with an image that looks close enough to someone without actually specifying in the prompt.
So unless people want to check every images for what prompts where used, we should probably just ban AI content outright.
So unless people want to check every images for what prompts where used, we should probably just ban AI content outright.
This was from the thread about RobbyWLP: https://umd.net/forums/producer-caught-selling-nonconsensual-ai?start=0#postid_1082859
dalamar666 said: We should not be using AI for anything related to this. AI being allowed on here in my mind is insulting to producers and models.
He's not outright saying to ban AI here, but the implication sounds pretty clear.
CreamMeAgain said: There's nothing wrong with creating a fantasy if AI is used respectfully. But that's the biggest glaring problem with AI, the technology itself isn't as dangerous as the people using it. And that's where I oppose the use of it here on this website. There is an epidemic in this community of piracy, and it's those same people who are looking at AI to make stealing easier for them.
To be honest, I'm not sure if she opposes AI in general on here, or just its use for piracy and using real people's likenesses. To be clear, I agree with her comment except for the opposition of it on this website.
Kabe22 said: I'm actually in favor of:
Requiring all AI content to be tagged as such (with the exception AI used in the editing of real videos/photosets, within reason)
The AI filter, so those who don't want to see it don't have to
AI content in the forums being limited to the AI WAM group or maybe a more dedicated subforum if AI gains more traction on this site.
A complete ban on AI images/video/etc of real people without the subject's consent
A complete ban of copyright-infringing AI
A ban on links to off-site media that violate either of those other bans.
As someone who posts exclusively AI images/gifs, I have yet to have site admin/mods block/deny anything I've posted, so I'm not concerned about them trying to ban AI.
Requiring all AI content to be tagged as such (with the exception AI used in the editing of real videos/photosets, within reason) The AI filter, so those who don't want to see it don't have to AI content in the forums being limited to the AI WAM group or maybe a more dedicated subforum if AI gains more traction on this site. A complete ban on AI images/video/etc of real people without the subject's consent A complete ban of copyright-infringing AI A ban on links to off-site media that violate either of those other bans. As someone who posts exclusively AI images/gifs, I have yet to have site admin/mods block/deny anything I've posted, so I'm not concerned about them trying to ban AI.
Okay - I'm with you so far.
After re-reading both this thread and the one about Robby/Photoslop, I didn't find as many supporting a general ban as I remembered (most were talking about banning non-consentual AI, which I agree with). There were these, though (examples follow)
Good! So then let's look at these one at a time.
dalamar sure does seem to be against AI being allowed, although it's interesting that they don't come right out and say that the admins should ban it. I'd be curious to see if that's actually what they want or if they're just, like, venting or being hyperbolic or something. (As I said earlier, I usually don't like trying to read strangers' minds over the internet, no matter how obvious the implications may seem.)
Potato's point is a curious one, because it goes right back to something that you just said:
A complete ban on AI images/video/etc of real people without the subject's consent A complete ban of copyright-infringing AI Potato's stated view errs on the side of caution here (perhaps in an extreme way), and maybe some part of his background reasoning is off in some way, but it seems to me that he's using the same starting premise that you are (namely, no unconstenting real people or copyrighted stuff). So if you're going to cite him as someone who's, like, unfairly endangering the future of AI WAM or something, then it seems to me that it's important for you to say precisely where he went wrong. To me, his values seem to be the same as your values, so it's not immediately obvious to me how you two are reaching opposite conclusions on the subject.
The same, I think, goes with CreamMe. It's a little puzzling to me to see you say that you agree with what she says except for the part where she draws a conclusion that's different from yours. Like, I come into these conversations with the assumption that people who have (roughly) the same values and the same evidence should reach (roughly) the same conclusion. But it seems like that's not happening for you and these other people, and I'm not sure why.
(And, again, probably all of this is academic for the time being? Like you said, there's no real indication that anybody actually will ban AI here any time soon, so maybe we're just counting angels on pin-heads. It's just weird to me that so many people in this thread seem to be either talking past each other or saying similar things but in an antagonistic way.)
larryniven said: And, again, probably all of this is academic for the time being? Like you said, there's no real indication that anybody actually will ban AI here any time soon, so maybe we're just counting angels on pin-heads. It's just weird to me that so many people in this thread seem to be either talking past each other or saying similar things but in an antagonistic way.
There we are then
Sponsors
To avoid content being blocked due to your local laws, please verify your age ?
Sponsors
VIP Coupons
Design & Code ©1998-2025 Loverbuns, LLC 18 U.S.C. 2257 Record-Keeping Requirements Compliance Statement
Epoch Billing Support Log In