I get this and I don’t disagree, but I also hate that AI fully brought back thought crimes as a thing.
I don’t have a better approach or idea, but I really don’t like that simply drawing a certain arrangement of lines and colors is now a crime. I’ve also seen a lot of positive sentiment at applying this to other forms of porn as well, ones less universally hated.
Not supporting this use case at all and on balance I think this is the best option we have, but I do think thought crimes as a concept are just as concerning, especially given the current political climate.
I really don’t like that simply drawing a certain arrangement of lines and colors is now a crime
I’m sorry to break it to you, but this has been illegal for a long time and it doesn’t need to have anything to do with CSAM.
For instance, drawing certain copyrighted material in certain contexts can be illegal.
To go even further, numbers and maths can be illegal in the right circumstances. For instance, it may be illegal where you live to break the encryption of a certain file, depending on the file and encryption in question (e.g. DRM on copyrighted material). “Breaking the encryption of a file” essentially translates to “doing maths on a number” when you boil it down. That’s how you can end up with the concept of illegal numbers.
To further clarify it’s specifically around thought crimes in scenarios where there is no victim being harmed.
If I’m distributing copyrighted content, that’s harming the copyright holder.
I don’t actually agree with breaking DRM being illegal either, but at least in that case, doing so is supposedly harming the copyright holder because presumably you might then distribute it, or you didn’t purchase a second copy in the format you wanted or whatever. There’s a ‘victim’ that’s being harmed.
Doodling a dirty picture of a totally original character doing something obscene harms absolutely no one. No one was abused. No reputation (other than my own) was harmed. If I share that picture with other consenting adults in a safe fashion, again no one was harmed or had anything done to them that they didn’t agree to.
It’s totally ridiculous to outlaw that. It’s punishing someone for having a fantasy or thought that you don’t agree with and ruining their life. And that’s an extremely easy path to expand into other thoughts you don’t like as well. And then we’re back to stuff like sodomy laws and the like.
Sure, i think it’s weird to really care about loli or furry or any other niche the way a lot of people do around here, but ai generating material of actual children (and unwilling people besides) is actually harmful. If they can’t have effective safeguards against that harm it makes sense to restrict it legally.
Making porn of actual people without their consent regardless of age is not a thought crime. For children, that’s obviously fucked up. For adults it’s directly impacting their reputation. It’s not a victimless crime.
But generating images of adults that don’t exist? Or even clearly drawn images that aren’t even realistic? I’ve seen a lot of people (from both sides of the political spectrum) advocate that these should be illegal if the content is what they consider icky.
Like let’s take bestiality for example. Obviously gross and definitely illegal in real life. But should a cartoon drawing of the act really be illegal? No one was abused. No reputation was damaged. No illegal act took place. It was simply someone’s fucked up fantasy. Yet lots of people want to make that into a thought crime.
I’ve always thought that if there isn’t speech out there that makes you feel icky or gross then you don’t really have free speech at all. The way you keep free speech as a right necessarily requires you to sometimes fight for the right of others to say or draw or write stuff that you vehemently disagree with, but recognize as not actually causing harm to a real person.
GenAI is vastly different though. Those are known to sometimes regurgitate people or things from their dataset, (mostly) unaltered. Like how you can get Copilot to spit out valid secrets that people accidentally committed by typing NPM_KEY=. You can’t have any guarantee that if you ask it to generate a picture of a person, that person does not actually exist.
You can insist every frame of Bart Simspon’s dick in The Simpsons Movie should be as illegal as photographic evidence of child rape, but that does not make them the same thing. The entire point of the term CSAM is that it’s the actual real evidence of child rape. It is nonsensical to use the term for any other purpose.
“CSAM (“see-sam”) refers to any visual content—photos, videos, livestreams, or AI-generated images—that shows a child being sexually abused or exploited.”
“Any content that sexualizes or exploits a child for the viewer’s benefit” <- AI goes here.
RAINN has completely lost the plot by conflating the explicit term for Literal Photographic Evidence Of An Event Where A Child Was Raped with made-up bullshit.
We will inevitably develop some other term like LPEOAEWACWR, and confused idiots will inevitably misuse that to refer to drawings, and it will be the exact same shit I’m complaining about right now.
Child pornography (CP), also known as child sexual abuse material (CSAM) and by more informal terms such as kiddie porn, is erotic material that involves or depicts persons under the designated age of majority.
[…]
Laws regarding child pornography generally include sexual images involving prepubescents, pubescent, or post-pubescent minors and computer-generated images that appear to involve them.
(Emphasis mine)
The Rape, Abuse, & Incest National Network (RAINN) defines child sexual abuse material (CSAM) as “evidence of child sexual abuse” that "includes both real and synthetic content
You think these billion-dollar companies keep hyper-illegal images around, just to train their hideously expensive models to do the things they do not want those models to do?
Like combining unrelated concepts isn’t the whole fucking point?
True enough - but fortunately, there’s approximately zero such images readily-available on public websites, for obvious reasons. There certainly is not some well-labeled training set on par with all the images of Shrek.
Yes and they’ve been proven to do so. Meta (Facebook) recently made the news for pirating a bunch of ebooks to train its AI.
Anna’s Archive, a site associated with training AI, recently scraped some 99.9% of Spotify songs. They say at some point they will make torrents so the common people can download it, but for now they’re using it to teach AI to copy music. (Note: Spotify uses lower quality than other music currently available, so AA will offer nothing new if/when they ever do release these torrents.)
So, yes, that is exactly what they’re doing. They are training their models on all the data, not just all the legal data.
It’s big fucking news when those datasets contain, like, three JPEGs. Because even one such JPEG is an event where the FBI shows up and blasts the entire hard drive into shrapnel.
Y’all insisting there’s gotta be some clearly-labeled archive with a shitload of the most illegal images imaginable, in order for the robot that combines concepts to combine the concept of “child” and the concept of “naked,” are not taking yourselves seriously. You’re just shuffling cards to bolster a kneejerk feeling.
It literally can’t combine unrelated concepts though. Not too long ago there was the issue where one (Dall-E?) couldn’t make a picture of a full glass of wine because every glass of wine it had been trained on was half full, because that’s generally how we prefer to photograph wine. It has no concept of “full” the way actual intelligences do, so it couldn’t connect the dots. It had to be trained on actual full glasses of wine to gain the ability to produce them itself.
I’m saying it can’t combine clothed children and naked adults to make naked children. It doesn’t know what “naked” means. It can’t imagine what something might look like. It can only make naked children if it has been trained on them directly.
Nothing made-up is CSAM. That is the entire point of the term “CSAM.”
It’s like calling a horror movie murder.
It’s too hard to tell real CSAM from AI-generated CSAM. Safest to treat it all as CSAM.
I get this and I don’t disagree, but I also hate that AI fully brought back thought crimes as a thing.
I don’t have a better approach or idea, but I really don’t like that simply drawing a certain arrangement of lines and colors is now a crime. I’ve also seen a lot of positive sentiment at applying this to other forms of porn as well, ones less universally hated.
Not supporting this use case at all and on balance I think this is the best option we have, but I do think thought crimes as a concept are just as concerning, especially given the current political climate.
I’m sorry to break it to you, but this has been illegal for a long time and it doesn’t need to have anything to do with CSAM.
For instance, drawing certain copyrighted material in certain contexts can be illegal.
To go even further, numbers and maths can be illegal in the right circumstances. For instance, it may be illegal where you live to break the encryption of a certain file, depending on the file and encryption in question (e.g. DRM on copyrighted material). “Breaking the encryption of a file” essentially translates to “doing maths on a number” when you boil it down. That’s how you can end up with the concept of illegal numbers.
To further clarify it’s specifically around thought crimes in scenarios where there is no victim being harmed.
If I’m distributing copyrighted content, that’s harming the copyright holder.
I don’t actually agree with breaking DRM being illegal either, but at least in that case, doing so is supposedly harming the copyright holder because presumably you might then distribute it, or you didn’t purchase a second copy in the format you wanted or whatever. There’s a ‘victim’ that’s being harmed.
Doodling a dirty picture of a totally original character doing something obscene harms absolutely no one. No one was abused. No reputation (other than my own) was harmed. If I share that picture with other consenting adults in a safe fashion, again no one was harmed or had anything done to them that they didn’t agree to.
It’s totally ridiculous to outlaw that. It’s punishing someone for having a fantasy or thought that you don’t agree with and ruining their life. And that’s an extremely easy path to expand into other thoughts you don’t like as well. And then we’re back to stuff like sodomy laws and the like.
Sure, i think it’s weird to really care about loli or furry or any other niche the way a lot of people do around here, but ai generating material of actual children (and unwilling people besides) is actually harmful. If they can’t have effective safeguards against that harm it makes sense to restrict it legally.
Making porn of actual people without their consent regardless of age is not a thought crime. For children, that’s obviously fucked up. For adults it’s directly impacting their reputation. It’s not a victimless crime.
But generating images of adults that don’t exist? Or even clearly drawn images that aren’t even realistic? I’ve seen a lot of people (from both sides of the political spectrum) advocate that these should be illegal if the content is what they consider icky.
Like let’s take bestiality for example. Obviously gross and definitely illegal in real life. But should a cartoon drawing of the act really be illegal? No one was abused. No reputation was damaged. No illegal act took place. It was simply someone’s fucked up fantasy. Yet lots of people want to make that into a thought crime.
I’ve always thought that if there isn’t speech out there that makes you feel icky or gross then you don’t really have free speech at all. The way you keep free speech as a right necessarily requires you to sometimes fight for the right of others to say or draw or write stuff that you vehemently disagree with, but recognize as not actually causing harm to a real person.
Drawings are one conversation I won’t get into.
GenAI is vastly different though. Those are known to sometimes regurgitate people or things from their dataset, (mostly) unaltered. Like how you can get Copilot to spit out valid secrets that people accidentally committed by typing
NPM_KEY=. You can’t have any guarantee that if you ask it to generate a picture of a person, that person does not actually exist.You can insist every frame of Bart Simspon’s dick in The Simpsons Movie should be as illegal as photographic evidence of child rape, but that does not make them the same thing. The entire point of the term CSAM is that it’s the actual real evidence of child rape. It is nonsensical to use the term for any other purpose.
You are completely wrong.
https://rainn.org/get-the-facts-about-csam-child-sexual-abuse-material/what-is-csam/
“CSAM (“see-sam”) refers to any visual content—photos, videos, livestreams, or AI-generated images—that shows a child being sexually abused or exploited.”
“Any content that sexualizes or exploits a child for the viewer’s benefit” <- AI goes here.
RAINN has completely lost the plot by conflating the explicit term for Literal Photographic Evidence Of An Event Where A Child Was Raped with made-up bullshit.
We will inevitably develop some other term like LPEOAEWACWR, and confused idiots will inevitably misuse that to refer to drawings, and it will be the exact same shit I’m complaining about right now.
Dude, you’re the only one who uses that strict definition. Go nuts with your course of prescriptivism but I’m pretty sure it’s a lost cause.
https://en.wikipedia.org/wiki/Child_pornography
‘These several things are illegal, including the real thing and several made-up things.’
Please stop misusing the term that explicitly refers to the the real thing.
‘No.’
deleted by creator
Is it a sexualized depiction of a minor? Then it’s csam. Fuck all y’all pedo apologists.
Were you too busy fapping to read the article?
AI CSAM was generated from real CSAM
AI being able to accurately undress kids is a real issue in multiple ways
AI can draw Shrek on the moon.
Do you think it needed real images of that?
It used real images of shrek and the moon to do that. It didnt “invent” or “imagine” either.
The child porn it’s generating is based on literal child porn, if not itself just actual child porn.
You think these billion-dollar companies keep hyper-illegal images around, just to train their hideously expensive models to do the things they do not want those models to do?
Like combining unrelated concepts isn’t the whole fucking point?
No, I think these billion dollar companies are incredibly sloppy about curating the content they steal to train their systems on.
True enough - but fortunately, there’s approximately zero such images readily-available on public websites, for obvious reasons. There certainly is not some well-labeled training set on par with all the images of Shrek.
Yes and they’ve been proven to do so. Meta (Facebook) recently made the news for pirating a bunch of ebooks to train its AI.
Anna’s Archive, a site associated with training AI, recently scraped some 99.9% of Spotify songs. They say at some point they will make torrents so the common people can download it, but for now they’re using it to teach AI to copy music. (Note: Spotify uses lower quality than other music currently available, so AA will offer nothing new if/when they ever do release these torrents.)
So, yes, that is exactly what they’re doing. They are training their models on all the data, not just all the legal data.
It’s big fucking news when those datasets contain, like, three JPEGs. Because even one such JPEG is an event where the FBI shows up and blasts the entire hard drive into shrapnel.
Y’all insisting there’s gotta be some clearly-labeled archive with a shitload of the most illegal images imaginable, in order for the robot that combines concepts to combine the concept of “child” and the concept of “naked,” are not taking yourselves seriously. You’re just shuffling cards to bolster a kneejerk feeling.
It literally can’t combine unrelated concepts though. Not too long ago there was the issue where one (Dall-E?) couldn’t make a picture of a full glass of wine because every glass of wine it had been trained on was half full, because that’s generally how we prefer to photograph wine. It has no concept of “full” the way actual intelligences do, so it couldn’t connect the dots. It had to be trained on actual full glasses of wine to gain the ability to produce them itself.
And you think it’s short on images of fully naked women?
I’m saying it can’t combine clothed children and naked adults to make naked children. It doesn’t know what “naked” means. It can’t imagine what something might look like. It can only make naked children if it has been trained on them directly.
Incorrect.