US man used AI to generate 13,000 child sexual abuse pictures, FBI alleges
US man used AI to generate 13,000 child sexual abuse pictures, FBI alleges
Steven Anderegg allegedly used the Stable Diffusion AI model to generate photos; if convicted, he could face up to 70 years in prison
cross-posted from: https://lemmy.zip/post/15863526
Steven Anderegg allegedly used the Stable Diffusion AI model to generate photos; if convicted, he could face up to 70 years in prison
You're viewing a single thread.
How are they abuse images if no abuse took place to create them?
Because they are images of children being graphically raped, a form of abuse. Is an AI generated picture of a tree not a picture of a tree?
No it isn't, not anymore than a drawing of a car is a real car, or drawings of money are real money.
Material showing a child being sexually abused is child sexual abuse material.
And an AI generated image does not show a child being abused
Bullshit and you know it.
There is no child being abused by a generated image or a drawing.
If Paedophile Hill is the hill you want to die on, it's no loss to me, so I've got zero interest in your "Ceci n'est pas une child rape" defense.
And yet you still engaged with it. If we're gonna classify every picture/drawing/gen that makes people uncomfortable as CSAM it distracts from the actual CSAM that is running rampant
Seems alot like your just promoting CSAM at that point.
Im sure that will have absolutely no effect on the pedophiles who are attracted to children acting on their desires. /s
It distracts from the very real issue of actual CSAM being produced and spread by pieces of shit like the guy in the article. Instead, people are going to focus on the guy using stable diffusion to generate images that are not the result of actual abuse. Its exactly what companies like OpenAI and Microsoft want, demonization of the open source AI projects. And you and many others in this thread are falling for it.
I'm not engaging for your benefit, which is why I've got no interest in repeating the same point in 500 ways in the hope it sinks in. But the reality is that a lot of people get their opinions from social media and they sure as fuck shouldn't imitate your views on CSAM so it's important that nobody mistakes contrarianism and apologism for actual wisdom.
But yes, it is hard to stand by while you lie your little heart in a way that helps paedophiles. I'm not ashamed or embarrassed about that.
So here's how it will play out: Your bullshit apologism and enabling will result in the creation of platforms for circulating child pornography. This platform will immediately be flooded with pictures and videos of children being raped that are indistinguishable from "genuine" child pornography, thanks to models being trained on paedophiles back catalogue.
As the amount of content grows, more and more videos of actual children being raped will enter circulation, with moderators and paedophile wriggling out of it by claiming "I thought it was AI generated".
New videos featuring the rape of actual children will be created and posted to these communities as child pornography normalises the abuse of children for the members. Detection and prosecution of the people responsible being functionally impossible because they've been buried and obfuscated by the AI generated content you insist doesn't count.
But hey, at least your bullshit semantic sensibilities haven't been offended right? That seems way more important to you than the abuse of children anyway. You're basically a hero for selflessly safeguarding paedophiles jerk off material.
We're not talking about "drawings of children being raped that make people uncomfortable". We're talking about pictures and videos that are indistinguishable from reality, featuring children being coereced or forced into performing every act and fetish known to pornography.
And you fucking know it.
Nobody is saying they're real, and I now see what you're saying.
By your answers, your question is more "at-face-value" than people assume:
You are asking:
"Did violence occur in real life in order to produce this violent picture?"
The answer is, of course, no.
But people are interpreting it as:
"This is a picture of a man being stoned to death. Is this picture violent, if no violence took place in real life?"
To which answer is, yes.
It can be abhorrent and unlikable, its still not abuse
We're not disagreeing.
The question was:
"Is this an abuse image if it was generated?"
Yes, it is an abuse image.
Is it actual abuse? Of course not.
And yet its being treated as though it is
Images of children being raped are being treated as images of children being raped. Nobody has every been caught with child pornography and charged as if they abused the children themselves, nor is anybody advocating that people generating AI child pornography are charged as if they sexually abused a child.
Everything is being treated as it always has been, but you're here arguing that it's moral and harmless as long as an AI does it, using every semantic trick and shifted goalpost you possibly can.
It's been gross as fuck to watch. I know you're aiming for a kind of "king of rationality, capable of transcending even your disgust of child abuse" thing, but every argument you make is so trivial and unimportant that you're coming across as someone hoping CSAM becomes more accessible.
Well, that's another story. I just answered your question. "Are these images about abuse even if they're generated?" Yup, they are.
"Should people be prosecuted because of them?" Welp, someone with more expertise should answer this. Not me.
No genius it's just promoting abuse. Have a good day.
Just like violent video games produce school shooters
You've already fucked up your own argument. You're supposed to be insisting there's no such thing as a "violent video game", because representations of violence don't count, only violence done to actual people.
If you can't follow a simple line of logic to explain a counterpoint, that's on you.
I understood it just fine.
Oops you forgot to use logic. As per the comment you're replying to, the more apt analogy would be: is an AI generated picture of a car still a picture of a car.
That has nothing to do with logic? Its pointing out that both drawings and AI gens are not really the things they might depict
It's a picture of a hallucination of a tree. Distinguishing real from unreal ought to be taken more seriously given the direction technology is moving.
It's a picture of a hallucination of a tree
So yes, it's a tree. It's a tree that might not exist, but it's still a picture of a tree.
You can't have an image of a child being raped -- regardless of if that child exists or not -- that is not CSAM because it's an image of a child being sexually abused.
Distinguishing real from unreal ought to be taken more seriously given the direction technology is moving.
Okay, so who are you volunteering to go through an endless stream of images and videos of children being raped to verify that each one has been generated by an AI and not a scumbag with a camera? Peados?
Why are neckbeards so enthusiastic about dying on this hill? They seem more upset that there's something they're not allowed to jerk off to than by the actual abuse of children.
Functionally, legalising AI generated CSAM means legalising "genuine" CSAM because it will be impossible to distinguish the two, especially as paedophiles dump their pre-AI collections or feed them in as training data.
People who do this are reprehensible, no matter what hair splitting and semantic gymnastics they employ.
Hey man, I'm not the one. I'm literally just saying that the images that AI creates are not real. If you're going to argue that they are, you're simply wrong. Should these ones be generated? Obviously I'd prefer that they not be. But they're still effectively fabrications that I'm better off simply not knowing about.
If you want to get into the weeds and discuss the logistics of enforcing what is essentially thought crime, that is a different discussion I'm frankly not savvy enough to have here. I have no control over the ultimate outcome, but for what it's worth, my money says thought crime will in fact become a punishable offense within our lifetimes, and this may well be an easy catalyst to use to that end. This should put your mind at ease.
The thread is about "how are they abuse images if no abuse took place" and the answer is "because they're images of abuse". I haven't claimed they're real at any point.
It's not a thought crime because it's not a thought. Nobody is being charged for thinking about raping children, they're being charged for creating images of children being raped.
If the images are generated and held by a single person, it may as well be a thought crime. If I draw a picture of a man killing an animal, which is an image depicting a heinous crime spawned by my imagination, and I go to prison over this image, I would consider this a crime of incorrect thought. There are no victims, no animals are harmed, but my will spawned an image of a harmed animal. Authorities dictated I am not allowed to imagine this scenario. I am punished for it. I understand that the expression of said thought is what's being punished, but that is very literally the only way to punish a thought to begin with (for now), hence freedom of expression being a protected right.
The reason this is a hard issue to discuss in this context is because the topic at hand is visceral and charged. No one wants to be caught dead defending the rights of a monster, lest they be labeled a monster themselves. I see this as a failure of society to know what to do about people like this, opting instead to throw them into a box and hope they die there. If our justice system wasn't so broken, I might give less of a shit, but as it stands I see this response as shortsighted and inhumane.
If the model was trained on csam then it is dependent on abuse
That's a heck of a slippery slope I just fell down.
If responses generated from AI can be held criminally liable for their training data's crimes, we can all be held liable for all text responses from GPT, since it's being trained on reddit data and likely has access to multiple instances of brigading, swatting, man hunts, etc.
You just summarized the ongoing ethical concerns experts and common folk alike have been talking about in the past few years.
As I said in my other comment, the model does not have to be trained on CSAM to create images like this.
That irrelevant, any realistic depiction of children engaged in sexual activity meets the legal definition of csam. Even using filters on images of consenting adults could qualify as csam if the intent was to make the actors appear underage.
doesn't even have to be that realistic.
All the lemmy.world commenters came out to insist "that painting is a pipe, though."
Yeah? Smoke it.
Lemmy.world and bandwagoning on a sensitive topic that they know nothing about? Classic combo.
You'd figure "CSAM" was clear enough. You'd really figure. But apparently we could specify "PECR" for "photographic evidence of child rape" and people would still insist "he drew PECR!" Nope. Can't. Try again.
Ever moving goal posts. Ever notice how the ones who cry "for the children" the most seemingly have the most to hide?
I mean... regardless of your moral point of view, you should be able to answer that yourself. Here's an analogy: suppose I draw a picture of a man murdering a dog. It's an animal abuse image, even though no actual animal abuse took place.
Its not though, its just a drawing.
Except that it is an animal abuse image, drawing, painting, fiddle, whatever you want to call it. It's still the depiction of animal abuse.
Same with child abuse, rape, torture, killing or beating.
Now, I know what you mean by your question. You're trying to establish that the image/drawing/painting/scribble is harmless because no actual living being suffering happened. But that doesn't mean that they don't depict it.
Again, I'm seeing this from a very practical point of view. However you see these images through the lens of your own morals or points of view, that's a totally different thing.
And when characters are killed on screen in movies, are those snuff films?
No, they're violent films.
Snuff is a different thing, because it's supposed to be real. Snuff films depict violence in a very real sense. So so they're violent. Fiction films also depict violence. And so they're violent too. It's just that they're not about real violence.
I guess what you're really trying to say is that "Generated abuse images are not real abuse images." I agree with that.
But at face value, "Generated abuse images are not abuse images" is incorrect.
AI does not create content from nothing. It must be trained by being fed examples to generate data from. An AI cannot generate CSAM without being fed CSAM. Therefore, yes, abuse HAS taken place in order to create it.
CSAM is illegal all around
It isn't csam if there was no abuse.
It's not child sexual assault if there was no abuse. However, the legal definition of csam is any visual depiction, including computer or computer-generated images of sexually explicit conduct, where [...]— (A) the production of such visual depiction involves the use of a minor engaging in sexually explicit conduct; (B) such visual depiction is a digital image, computer image, or computer-generated image that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct; or (C) such visual depiction has been created, adapted, or modified to appear that an identifiable minor is engaging in sexually explicit conduct.
You may not agree with that definition, but even simulated images that look like kids engaging in sexual activity meet the threshold for CSAM.
Yes it is
CSAM is child pornography
Do you not know that CSAM is an acronym that stands for child sexual abuse material?
True but CSAM is anything that involves minors. Its really up to the court to decide a lot of it but in the case above I'd imagine that the images were quite disturbing.
in this instance, no human children or minors of any kind were involved.
I think the court looked at the phycological aspects of it. When you look at that kind of material you are training your brain and body to be attracted to that stuff in real life.
prove that any "training" is involved, please.
We're discussing the underpinnings and philosophy of the legality and your comment is simply "it is illegal"
I can only draw from this that your morality is based on laws instead of vice versa.
I'm in the camp if that there is no reason that you should have that kind of imagery especially AI generated imagery. Think about what people often do with pornography. You do not want them doing that with children regardless of if it is AI generated.
What does want have to do with it? I'd rather trust science and psychologists to determine if this, which is objectively harmless, helps them control their feelings and gives them a harmless outlet.
They aren't banning porn in general. They just don't want to create any more sexual desires toward children. The CSAM laws came from child protection experts. Admittedly some of these people want to "ban" encryption but that's irrelevant in this case.
There was no C.
There was no SA.
The entire point of saying "CSAM" was to distinguish evidence of child rape from depictions of imaginary events.
While I can't say for 100% certain, it would almost assuradly be trained on many abuse images in order to generate more.
No that's not necessarily true at all. You can take a stock stable diffusion model and output these images right Now without additional training. The whole point of diffusion models is that you don't need 100% training data coverage to create new images outside of the original dataset. Having learned the concept of child from any normal image set of children and learning the concept of nudity/porn from legal adult images is more than enough to create a blended concept of the two.
Having learned the concept of child from any normal image set of children
Those children are the abuse victims.
Having never used an AI generator, generic generated images wouldn't be an actual match to the dataset images, right? It would just be generating features it understands to be associated with the concept of a child, which would make the claim that the dataset children are the abuse targets a stretch, unless there's some other direct or indirect harm to them. An immediate exception being a person writing a prompt attempting to create a specific facsimile of an individual.
Section 1466A of Title 18, United States Code, makes it illegal for any person to knowingly produce, distribute, receive, or possess with intent to transfer or distribute visual representations, such as drawings, cartoons, or paintings that appear to depict minors engaged in sexually explicit conduct and are deemed obscene.
That's nice, still illegal.
While true, it also wasn't really what my post was responding to. Thanks though.
The thread is discussing why it's considered abuse if you can't point to a victim. The answer turned out to be "because the law says so."
If you read the law you posted, it doesn't actually address the question of victimhood. Also, I don't really get why you're still trying to force an unrelated point into this part of the discussion. Maybe find another place in the thread where someone thinks it's legal and go talk to them.
The children don't exist. The concept of child is learned from a series of 1s and 0s
Then it's abuse because the law says so and how the images are generated is irrelevant.
Believe it or not, US law isn't the moral center of the universe not to mention its disproportionate use against groups it doesn't like.
This case is in the US.
Yeah and he should go to jail for grooming a kid but to try and pin csam creation charges on him is wrong, regardless of the US' fucked up laws.