How would you respond to photo realistic porn that looks like your mother, daughter, [insert person you care about here] especially if they found it distressing?
How would you feel if it was posted on facebook? How would you feel if they had to deal with it at work? From coworkers? From clients?
We are entering uncharted waters. You know why this is different than training a model on text, and your reply to @GBU_28@lemm.ee is hostile and doesn’t acknowledge why people would be upset about AI porn featuring their likeness.
you are answering a question with a different question. LLMs don’t make pictures of your mom. And this particular question?. One that has roughly existed since Photoshop existed.
It just gets easier every year. It was already easy. You could already pay someone 15 bucks on Fiver to do all of that, for years now.
Nothing really new here.
The technology is also easy. Matrix math. About as easy to ban as mp3 downloads. Never stopped anyone. It’s progress. You are a medieval knight asking to put gunpowder back into the box, but it’s clear it cannot be put back - it is already illegal to make non consensual imagery just as it is illegal to copy books. And yet printers exist and photocopiers exist.
Let me be very clear - accepting the reality that the technology is out there, it’s basic, easy to replicate and on a million computers now is not disrespectful to victims of no consensual imagery.
You may not want to hear it, but just like with encryption, the only other choice society has is full surveillance of every computer to prevent people from doing “bad things”. everything you complain about is already illegal and has already been possible - it just gets cheaper every year. What you want to have protection from is technological progress because society sucks at dealing with the consequences of it.
To be perfectly blunt, you don’t need to train any generative AI model for powerful deepfakes. You can use technology like Roop and Controlnet to synthesize any face on any image from a singe photograph. Training not necessary.
When you look at it that way, what point is there to try to legislate training with these arguments? None.
I’m not making an argument to ban it. I’m just pointing out you’re pretending a model from text someone wrote is similar to a model that makes nonconcentual porn.
I don’t think it can be banned, it’s just something they will need to encorperate into revenge porn laws, if it isn’t already covered.
Oh good, someone who has read every revenge porn law, ever. I’m glad they work exactly the same, in every nation and state.
Anyway, I must be hallucinating, true, because it seems you keep attacking what I’m saying, instead of defending the comment you made earlier that I took issue with, the one that points out you’re being needlessly hostile.
I can do this right now with photoshop dude what are you talking about. This just points to the need for more revenge porn laws.
We don’t have to sit in the fire when we can crawl out. Are we still on fire? Yeah. Can we do something about that? Yeah!
It seems like so many people these days want perfect solutions but the reality is that sometimes we have to make incremental solutions to erase the problem as much as we can.
And incidentally, this need for revenge porn laws is also a symptomatic issue with a separate cause. Technology always moved forward and with no relation to social advancement, where there is also no realistic “Genie being forced back in the bottle” scenario either.
That being said, easier access to more powerful technology with lackluster recognition of personal responsibility doesn’t exactly bring happy prospects. lol…
Revenge porn/blackmail/exploitation will hopefully become much less obscene, not to the “let’s not prosecute this” levels, but maybe people can stop living in fear of their lives being ruined by malicious actors (unless that’s your kink, you do you).
It will take/drive/demand a massive cultural shift, but if you asked me which world I would rather live in, and the options are one where people are abused and exploited, or one where people can visualize their perversions more easily (but content creators have a harder time making a living) I’ll take the second. Though I may have straw-manned a bit, it’s not something I’ve thought of outside of this forum thread.
The jist is that LLM find similar “chunks” out content from their training set, and assemble a response based on this similarity score (similar to your prompt request).
They know nothing they haven’t seen before, and the nicely of them is they create new things from parts of their training data.
Obviously they are very impressive tools but the concern is you can easily take a model that’s designed for porn, feed it pictures of someone you want to shame, and have it generate lifelike porn of a non porn actor.
That, and the line around “ethical” AI porn is blurry.
Strictly speaking, you arguably don’t either. Your knowledge of the world is based on your past experiences.
You do have more-sophisticated models than current generative AIs do, though, to construct things out of aspects of the world that you have experienced before.
The current crop are effectively more-sophisticated than simply pasting together content – try making an image and then adding “blue hair” or something, and you can get the same hair, but recolored. And they ability to replicate artistic styles is based on commonalities in seen works, but you don’t wind up seeing chunks of material just done by that artist.
Like, you have a concept of relative characteristics, and the current generative AIs do not. You can tell a human artist “make those breasts bigger”, and they can extrapolate from a model built on things they’ve seen before. The current crop of generative AIs cannot. But I expect that the first bigger-breast generative AI is going to attract users, based on a lot of what generative AIs are being used for now.
There is also, as I understand it, some understanding of depth in images in some existing systems, but the current generative AIs don’t have a full 3d model of what they are rendering.
But they’ll get more-sophisticated.
I would imagine that there will be a combination of techniques. LLMs may be used, but I doubt that they will be pure LLMs.
Ok, you know it’s trained on existing imagery right?
Sure the net new photos aren’t net new abuses, but whatever abuses went into the training set are literally represented in the product.
To be clear I’m not fully porn shaming here, but I wanted to clarify that these tools are informed from something already existing and cant be fully separated from the training data.
You do know how LLM are trained right?
I just retained an LLM on your comment you put on the public internet. You feel violated enough to equate it to physical violation?
Why would I? Folks who have had real nudes of them posted on the Internet haven’t felt “physical violation” but they’ve certainly been violated.
If you had photos of me and trained a porn generating LLM on my photos and shared porn of me, in an identifiable way, I would consider that violation.
But simply taking my words in that simple sentence isn’t identifiable, unique, or revealing. So no.
Further, the original point was about the ethics of AI porn. You can’t get something from nothing.
How would you respond to photo realistic porn that looks like your mother, daughter, [insert person you care about here] especially if they found it distressing?
How would you feel if it was posted on facebook? How would you feel if they had to deal with it at work? From coworkers? From clients?
We are entering uncharted waters. You know why this is different than training a model on text, and your reply to @GBU_28@lemm.ee is hostile and doesn’t acknowledge why people would be upset about AI porn featuring their likeness.
you are answering a question with a different question. LLMs don’t make pictures of your mom. And this particular question?. One that has roughly existed since Photoshop existed.
It just gets easier every year. It was already easy. You could already pay someone 15 bucks on Fiver to do all of that, for years now.
Nothing really new here.
The technology is also easy. Matrix math. About as easy to ban as mp3 downloads. Never stopped anyone. It’s progress. You are a medieval knight asking to put gunpowder back into the box, but it’s clear it cannot be put back - it is already illegal to make non consensual imagery just as it is illegal to copy books. And yet printers exist and photocopiers exist.
Let me be very clear - accepting the reality that the technology is out there, it’s basic, easy to replicate and on a million computers now is not disrespectful to victims of no consensual imagery.
You may not want to hear it, but just like with encryption, the only other choice society has is full surveillance of every computer to prevent people from doing “bad things”. everything you complain about is already illegal and has already been possible - it just gets cheaper every year. What you want to have protection from is technological progress because society sucks at dealing with the consequences of it.
To be perfectly blunt, you don’t need to train any generative AI model for powerful deepfakes. You can use technology like Roop and Controlnet to synthesize any face on any image from a singe photograph. Training not necessary.
When you look at it that way, what point is there to try to legislate training with these arguments? None.
I’m not making an argument to ban it. I’m just pointing out you’re pretending a model from text someone wrote is similar to a model that makes nonconcentual porn.
I don’t think it can be banned, it’s just something they will need to encorperate into revenge porn laws, if it isn’t already covered.
I’m just pointing out your comment sucked.
It’s already covered under those laws. So what are you doing that’s different from ChatGPT hallucinating here ?
Those laws don’t spell out the tools (photoshop); they hinge on reproducing likeness.
Oh good, someone who has read every revenge porn law, ever. I’m glad they work exactly the same, in every nation and state.
Anyway, I must be hallucinating, true, because it seems you keep attacking what I’m saying, instead of defending the comment you made earlier that I took issue with, the one that points out you’re being needlessly hostile.
I can do this right now with photoshop dude what are you talking about. This just points to the need for more revenge porn laws.
We don’t have to sit in the fire when we can crawl out. Are we still on fire? Yeah. Can we do something about that? Yeah!
It seems like so many people these days want perfect solutions but the reality is that sometimes we have to make incremental solutions to erase the problem as much as we can.
And incidentally, this need for revenge porn laws is also a symptomatic issue with a separate cause. Technology always moved forward and with no relation to social advancement, where there is also no realistic “Genie being forced back in the bottle” scenario either.
That being said, easier access to more powerful technology with lackluster recognition of personal responsibility doesn’t exactly bring happy prospects. lol…
Agreed, personal responsibility went out the window a long time ago. Apathy reigns supreme.
Revenge porn/blackmail/exploitation will hopefully become much less obscene, not to the “let’s not prosecute this” levels, but maybe people can stop living in fear of their lives being ruined by malicious actors (unless that’s your kink, you do you).
It will take/drive/demand a massive cultural shift, but if you asked me which world I would rather live in, and the options are one where people are abused and exploited, or one where people can visualize their perversions more easily (but content creators have a harder time making a living) I’ll take the second. Though I may have straw-manned a bit, it’s not something I’ve thought of outside of this forum thread.
Removed by mod
No so go ahead
Are you actually asking?
The jist is that LLM find similar “chunks” out content from their training set, and assemble a response based on this similarity score (similar to your prompt request).
They know nothing they haven’t seen before, and the nicely of them is they create new things from parts of their training data.
Obviously they are very impressive tools but the concern is you can easily take a model that’s designed for porn, feed it pictures of someone you want to shame, and have it generate lifelike porn of a non porn actor.
That, and the line around “ethical” AI porn is blurry.
Strictly speaking, you arguably don’t either. Your knowledge of the world is based on your past experiences.
You do have more-sophisticated models than current generative AIs do, though, to construct things out of aspects of the world that you have experienced before.
The current crop are effectively more-sophisticated than simply pasting together content – try making an image and then adding “blue hair” or something, and you can get the same hair, but recolored. And they ability to replicate artistic styles is based on commonalities in seen works, but you don’t wind up seeing chunks of material just done by that artist.
Like, you have a concept of relative characteristics, and the current generative AIs do not. You can tell a human artist “make those breasts bigger”, and they can extrapolate from a model built on things they’ve seen before. The current crop of generative AIs cannot. But I expect that the first bigger-breast generative AI is going to attract users, based on a lot of what generative AIs are being used for now.
There is also, as I understand it, some understanding of depth in images in some existing systems, but the current generative AIs don’t have a full 3d model of what they are rendering.
But they’ll get more-sophisticated.
I would imagine that there will be a combination of techniques. LLMs may be used, but I doubt that they will be pure LLMs.
deleted by creator
Ok, you know it’s trained on existing imagery right?
Sure the net new photos aren’t net new abuses, but whatever abuses went into the training set are literally represented in the product.
To be clear I’m not fully porn shaming here, but I wanted to clarify that these tools are informed from something already existing and cant be fully separated from the training data.