Alibaba Releases Advanced Open Video Model, Immediately Becomes AI Porn Machine
Alibaba Releases Advanced Open Video Model, Immediately Becomes AI Porn Machine
First off, I am sex positive, pro porn, pro sex work, and don't believe sex work should be shameful, and that there is nothing wrong about buying intimacy from a willing seller.
That said. The current state of the industry and the conditions for many professionals raises serious ethical issues. Coercion being the biggest issue.
I am torn about AI porn. On one hand it can produce porn without suffering, on the other hand it might be trained on other peoples work and take peoples jobs.
I think another major point to consider going forward is if it is problematic if people can generate all sorts of illegal stuff. If it is AI generated it is a victimless crime, so should it be illegal? I personally feel uncomfortable with the thought of several things being legal, but I can't logically argue for it being illegal without a victim.
I think another major point to consider going forward is if it is problematic if people can generate all sorts of illegal stuff. If it is AI generated it is a victimless crime, so should it be illegal? I personally feel uncomfortable with the thought of several things being legal, but I can't logically argue for it being illegal without a victim.
I've been thinking about this recently too, and I have similar feelings.
I'm just gonna come out and say it without beating around the bush: what is the law's position on AI-generated child porn?
More importantly, what should it be?
It goes without saying that the training data absolutely should not contain CP, for reasons that should be obvious to anybody. But what if it wasn't?
If we're basing the law on pragmatism rather than emotional reaction, I guess it comes down to whether creating this material would embolden paedophiles and lead to more predatory behaviour (i.e. increasing demand), or whether it would satisfy their desires enough to cause a substantial drop in predatory behaviour (I.e. lowering demand).
And to know that, we'd need extensive and extremely controversial studies. Beyond that, even in the event allowing this stuff to be generated is an overall positive (and I don't know whether it would or won't), will many politicians actually call for this stuff to be allowed? Seems like the kind of thing that could ruin a political career. Nobody's touching that with a ten foot pole.
It’s so much simpler than that—it can be created now, so it will be. They will use narrative twists to post it on the clearnet, just like they do with anime (she’s really a 1000 year old vampire, etc.). Creating laws to allow it are simply setting the rules of the phenomenon that is already going to be happening.
The only question is whether or not politicians will stop mud slinging long enough to have an adult conversation, or will we just shove everything into the more obscure parts of the internet and let it police itself.
No adult conversation required, just a quick "looks like we don't get internet privacy after all everyone." And erosion of more civil liberties. Again.
what is the law's position on AI-generated child porn?
Already illegal here in the UK https://metro.co.uk/2025/02/02/makers-ai-child-abuse-images-jailed-uk-introduces-world-first-law-22481459/
Illegal is most of the west already as creating sexual assault material of minors is already illegal regardless of method.
I think the concern is that although it's victimless, if it's legal it could.... Normalise (within certain circles) the practice. This might make the users more confident to do something that does create a victim.
Additionally, how do you tell if it's really or generated? If AI does get better, how do you tell?
what is the law’s position on AI-generated child porn?
Pretend underage porn is illegal in the EU and some other countries. I believe, in the US it is protected by the first amendment.
Mind that when people talk about child porn or CSAM that means anything underage, as far as politics is concerned. When two 17-year-olds exchange nude selfies, that is child porn. There were some publicized cases of teens in the US being convicted as pedophile sex offenders for sexting.
I believe, in the US it is protected by the first amendment.
CSAM, artificial or not, is illegal in the United States.
I see. I've looked up the details. Obscenity - whatever that means - is not protected by the first amendment. So where the material is obscene, it is still illegal.
https://en.wikipedia.org/wiki/Ashcroft_v._Free_Speech_Coalition
what is the law’s position on AI-generated child porn?
the simplest possible explanation here, is that any porn created based on images of children, is de facto illegal. If it's trained on adults explicitly, and you prompt it for child porn, that's a grey area, probably going to follow precedent for drawn art, rather than real content.
Let's play devils advocate. You find Bob the pedophile with pictures depicting horrible things. 2 things are true.
A third factor is that this technology and the inherent lack of privacy on the internet could potentially pinpoint numerous unknown pedophiles who can even if they haven't done any harm yet be profitably persecuted to societies ultimate profit so long as you value innocent kids more than perverts.
Am I reading this right? You're for prosecuting people who have broken no laws?
I'll add this; I have sexual fantasies (not involving children) that would be repugnant to me IRL. Should I be in jail for having those fantasies, even though I would never act on them?
This sounds like some Minority Report hellscape society.
Correct. This quickly approaches thought crime.
What about an AI gen of a violent rape and murder. Shouldn't that also be illegal.
But we have movies that have protected that sort of thing for years; graphically. Do those the become illegal after the fact?
And we also have movies of children being victimized so do these likewise become illegal?
We already have studies that show watching violence does not make one violent and while some refuse to accept that, it is well established science.
There is no reason to believe the same isn't true for watching sexual assault. There are been many many movies that contain such scenes.
But ultimately the issue will become that there is no way to prevent it. The hardware to generate this stuff is already in our pockets. It may not be efficient but it's possible and efficiency will increase.
The prompts to generate this stuff are easily shared and there is no way to stop that without monitoring all communication and even then I'm sure work around would occur.
Prohibition requires society sacrifice freedoms and we have to decide what weee willing to sacrifice here because as we've seen with or prohibitions, once we unleash the law on one, it can be impossible to undo.
Ok watch adult porn then watch a movie in which women or children are abused. Note how the abuse is in no way sexualized exactly opposite of porn. It often likely takes place off screen and when rape in general appears on screen between zero and no nudity co-occurs. For children it basically always happens off screen.
Simulated child abuse has been federally illegal for ~20 years in the US and we appear to have very little trouble telling the difference between prosecuting pedos and cinema even whilst we have struggled enough with sexuality in general.
But ultimately the issue will become that there is no way to prevent it.
This argument works well enough for actual child porn. We certainly don't catch it all but every prosecution takes one more pedo off the streets. The net effect is positive. We don't catch most car thieves either and nobody suggests we legalize car theft.
Am I reading this right? You’re for prosecuting people who have broken no laws?
No I'm for making it against the law to simulate pedophile shit as the net effect is fewer abused kids than if such images were to be legal. Notably you are free to fantasize about whatever you like its the actual creation and sharing of images that would be illegal. Far from being a minority report hellscape its literally the present way things already are many places.
Lol, how can you say that do confidently? How would you know that with fewer AI CP you get less abused kids? And what is the logic behind it?
Demand doesn’t really drop if something is illegal (same goes for drugs). The only thing you reduce is offering, which just resulting in making the thing that got illegal more valuable (this wakes attention of shady money grabbers that hate regulation / give a shit about law enforcement and therefore do illegal stuff to get money) and that you have to pay a shitton of government money maintaining all the prisons.
Basically every pedo in prison is one who isn't abusing kids. Every pedo on a list is one who won't be left alone with a young family member. Actually reducing AI CP doesn't actually by itself do anything.
Wrong. Every pedo in prison is one WHO HAS ALREADY ABUSED A CHILD, whether directly or indirectly. There is an argument to be made, and some studies that show, that dealing with Minor Attracted People before they cross the line can be effective. Unfortunately, to do this we need to be able to have a logical and civil conversation about the topic, and the current political climate does not allow for that conversation to be had. The consequence is that preventable crimes are not being prevented, and more children are suffering for it in the long run.
People are locked up all the time for just possessing child porn without having abused anyone. This isn't a bad thing because they are a danger to society.
No, they are not locked up because they're a danger to society. They're locked up because possessing CP is indirectly contributing to the abuse of the child involved.
Good arguments. I think I am convinced that both cases should be illegal.
If the pictures are real they probably increase demand, which is harmful. If the person knew, then the action therefore should result in jail and forced therapy.
If the pictures are not, forced therapy is probably the best option.
So I guess it being illegal in most cases simply to force therapy is the way to go. Even if it in one case is "victimless". If they don't manage to plausibly seem rehabilitated by professionals, then jail time for them.
I would assume (but don't really know) most pedophiles don't truly want to act on it, and don't want to have those urges. And would voluntarily go to therapy.
Which is why I am convinced prevention is the way to go. Not sacrificing privacy. In Norway we have anonymous ways for pedophiles to seek help. There have been posters and ads for it a lot of places a year back or something. I have not researched how it works in practice though.
Edit: I don't think the therapy we have in Norway is conversion therapy. It's about minimizing risk and helping deal with the underlying causes, medication, childhood trauma etc. I am not necessarily convinced that conversion therapy works.
Therapy is well and good and I think we need far more avenues available for people to get help (for all issues). That said, sexuality and attraction are complicated.
Let me start by saying I am not trying to state there is a 1:1 equivalence, this is just a comparison, but we have long abandoned conversion therapy for homosexuals, because we've found these preferences are core to them and not easily overwritten. The same is true for me as a straight person, I don't think therapy would help me find men attractive. I have to imagine the same is true for pedophiles.
The question is, if AI can produce pornography that can satisfy the urges of someone with pedophilia without harming any minors, is that a net positive? Remember the attraction is not the crime, it's the actions that harm others that are. Therapy should always be on the table.
This is a tricky subject because we don't want to become thought police, so all our laws are built in that manner. However there are big exceptions for sexual crimes due to the gravity of their impact on society. It's very hard to "stand up" for pedophilia because if acted upon it has monstrous effects, but AI is making us open this can of worms that I don't belive we ever really thought through besides criminalizing and demonizing (which could be argued was the correct approach with the technology at the time).
I really don't know enough about the subject or how that therapy works. I doubt that it is conversion therapy, but Ii really don't know. I would assume it's handling childhood trauma, medications etc.
Both therapy and if satisfying urges through AI generated content, is both something that should be answered scientifically. If there is research then that should be the basis for what decisions is taken, if there is a lack of research then more research should be the next step.
That said, sexuality and attraction are complicated.
There's nothing particularly complicated about it not being ok to rape kids, or to distribute depictions of kids being raped.
Nobody here has at all suggested it's ok to rape kids. I hope you can understand the difference between thinking something and doing something.
If the pictures are not, forced therapy is probably the best option.
This is true, but it really depends how "therapy" is defined. And forced therapy could mean anything from things like the old conversion therapy approach for gay people.
You might argue that these days we wouldn't do anything so barbaric, but considering that the nature of a pedophile's is very unsavory, and the fact that it will never be acceptable, unlike homosexuality, people would be far more willing to abuse or exploit said "therapy"
i have no problem with ai porn assuming it's not based on any real identities, i think that should be considered identity theft or impersonation or something.
Outside of that, it's more complicated, but i don't think it's a net negative, people will still thrive in the porn industry, it's been around since it's been possible, i don't see why it wouldn't continue.
Identity theft only makes sense for businesses. I can sketch naked Johny Depp in my sketchbook and do whatever I want with it and no one can stop me. Why should an AI tool be any different if distribution is not involved?
revenge porn, simple as. Creating fake revenge porn of real people is still to some degree revenge porn, and i would argue stealing someones identity/impersonation.
To be clear, you're example is a sketch of johnny depp, i'm talking about a video of a person that resembles the likeness of another person, where the entire video is manufactured. Those are fundamentally, two different things.
Again you're talking about distribution
I guess the point is this enables the mass production of revenge porn essentially at a person on the street level which makes it much harder to punish and prevent distribution. when it is relatively few sources that produces the unwanted product then only punishing the distribution might be a viable method. But when the production method becomes available to the masses then the only feasible control mechanism is to try to regulate the production method. It is all a matter of where is the most efficient position to put the bottle neck.
For instance when 3D printing allows people to produce automatic rifles in their homes "saying civil use of automatic rifles is illegal so that is fine" is useless.
I think that's a fair point and I wonder how will this effect the freedom of expression on the internet. If you can't find the distributor then it'll be really tough to get a handle of this.
On the other hand the sheer over abundance could simply break the entire value of revenge porn as in "nothing is real anyway so it doesn't matter" sort of thing which I hope would be the case. No one will be watching revenge porn cause they generate any porn they want in a heartbeat. Thats the ideal scenario anyway.
It is indeed a complicated problem with many intertwined variables, wouldn't wanna be in the shoes of policy makers (assuming that they actually are searching for an honest solution and not trying to turn this into profit lol).
For instance too much regulation on fields like this essentially would kill high quality open source AI tools and make most of them proprietary software leaving the field in the mercy of tech monopolies. This is probably what these monopolies want and they will surely try to push things this way to kill competition (talk about capitalism spurring competition and innovation!). They might even don the cloak of some of these bad actors to speed up the process. Given the possible application range of AI, this is probably even more dangerous than flooding the internet with revenge porn.
%100 freedom, no regulations will essentially lead to a mixed situation of creative and possibly ground breaking uses of the tech vs many bad actors using the tech for things like scamming, disinformation etc. how it will balance out on the long run is probably very hard to predict.
I think two things are clear, 1-both extremities are not ideal, 2- between the two extremities %100 freedom is still the better option (the former just exchanges many small bad actors for a couple giant bad actors and chokes any possible good outcomes).
Based on these starting with a solution closer to the "freedom edge" and improving it step by step based on results is probably the most sensible approach.
sort of. There are arguments that private ownership of these videos is also weird and shitty, however i think impersonation and identity theft are going to the two most broadly applicable instances of relevant law here. Otherwise i can see issues cropping up.
Other people do not have any inherent rights to your likeness, you should not simply be able to pretend to be someone else. That's considered identity theft/fraud when we do it with legally identifying papers, it's a similar case here i think.
But the thing is it's not a relevant law here at all as nothing is being distributed and no one is being harmed. Would you say the same thing if AI is not involved? Sure it can be creepy and weird and whatnot but it's not inhertly harmful or at least it's not obvious how it would be.
the only perceivable reason to create these videos is either for private consumption, in which case, who gives a fuck. Or for public distribution, otherwise you wouldn't create them. And you'd have to be a bit of a weird breed to create AI porn of specific people for private consumption.
If AI isn't involved, the same general principles would apply, except it might include more people now.
I've been thinking about this more and I think one interesting argument here is "toxic culture growth". As in even if the thing is not distributed it might grow undesired toxic cultures througu indirect exposures (like social media or forums discussions) even without the direct sharing.
I think this is slippery to the point of government mind control but maybe there's something valuable to research here either way.
I’ve been thinking about this more and I think one interesting argument here is “toxic culture growth”. As in even if the thing is not distributed it might grow undesired toxic cultures through indirect exposures (like social media or forums discussions) even without the direct sharing.
this is another big potential as well. Does it perpetuate cultural behaviors that you want to see in society at large? Similar things like this have resulted from misogyny and the relevant history of feminism.
It's a whole thing.
I think this is slippery to the point of government mind control but maybe there’s something valuable to research here either way.
i think there is probably a level of government regulation that is productive, i'm just curious about how we even define where that line starts and ends, because there is not really an explicitly clear point, unless you have solid external inferences to start from.
Honestly I'm quite happy with "social justice warrior" approach. Sure it's flawed and weak to manipulation for now but as a strategy for society to self correct its quite brilliant.
I'm optimistic society itself should be able to correct itself for this issue as well though considering the current climate the correction might be very chaotic.
i have no problem with ai porn assuming it’s not based on any real identities
With any model in use, currently, that is impossible to meet. All models are trained on real images.
With any model in use, currently, that is impossible to meet. All models are trained on real images.
yes but if i go to thispersondoesnotexist.com and generate a random person, is that going to resemble the likeness of any given real person close enough to perceptibly be them?
You are literally using the schizo argument right now. "If an artists creates a piece depicting no specific person, but his understanding of persons is based inherently on the facial structures of other people that he knows and recognizes, therefore he must be stealing their likeness"
No, the problem is a lack of consent of the person being used.
And now, being used to generate depictions of rape and CSAM.
yeah but like, legally, is this even a valid argument? Sure there is techically probably like 0.0001% of the average person being used in any given result of an AI generated image. I don't think that gives anyone explicit rights to that portion however.
That's like arguing that a photographer who captured you in a random photo in public that became super famous is now required to pay you royalties for being in that image, even though you are literally just a random fucking person.
You can argue about consent all you want, but at the end of the day if you're posting images of yourself online, you are consenting to other people looking at them, at a minimum. Arguably implicitly consenting to other people being able to use those images. (because you can't stop people from doing that, except for copyright, but that's not very strict in most cases)
And now, being used to generate depictions of rape and CSAM.
i dont see how this is even relevant, unless the person in question is a minor, a victim, or becoming a victim, otherwise it's no different than me editing an image of someone to make it look like they got shot in the face. Is that shitty? Sure. But i don't know of any laws that prevent you from doing that, unless it's explicitly to do with something like blackmail, extortion, or harassment.
The fundamental problem here is that you're in an extremely uphill position to even begin the argument of "well it's trained on people so therefore it uses the likeness of those people"
Does a facial structure recognition model use the likeness of other people? Even though it can detect any person that meets the requirements established by its training data? There is no suitable method to begin to breakdown at what point that persons likeness begins, and at what point it ends. it's simply an impossible task.
yeah but like, legally, is this even a valid argument?
Personally, legal is only what the law allows the wealthy to do, and provides punishments for the working class.
Morally, that's what you're doing when you use AI to generate CSAM. Its the same idea why we ban all pre-created CSAM, as well, because you are victimizing the person every single time.
i dont see how this is even relevant, unless the person in question is a minor, a victim, or becoming a victim,
It makes them a victim.
But i don’t know of any laws that prevent you from doing that, unless it’s explicitly to do with something like blackmail, extortion, or harassment.
The law exists to protect the ruling class while not binding them, and to bind the working class without protecting them.
Does a facial structure recognition model use the likeness of other people?
Yes.
Even though it can detect any person that meets the requirements established by its training data? There is no suitable method to begin to breakdown at what point that persons likeness begins, and at what point it ends. it’s simply an impossible task.
Exactly. So, without consent, it shouldn't be used. Periodt.
Personally, legal is only what the law allows the wealthy to do, and provides punishments for the working class.
if you have schizoprehnia, sure. Legal is what the law defines as ok. Whether or not people get charged for it is another thing. The question is "do you have the legal right to do it or not"
Morally, that’s what you’re doing when you use AI to generate CSAM. Its the same idea why we ban all pre-created CSAM, as well, because you are victimizing the person every single time.
legally, the reasoning behind this is because it's just extremely illegal, there are almost no if not zero, where it would be ok or reasonable, and therefore the moral framework tends to be developed around that. I don't necessarily agree with it always being victimization, because there are select instances where it just doesn't really make sense to consider it that, there are acts you commit that would make it victimization. However i like to subscribe to the philosophy that it is "abusive" material, and therefore innately wrong. Like blackmail, i find that to be a little bit more strict and conducive to that sort of definition.
It makes them a victim.
at one point in time yes, perpetually in some capacity, they will exist as having been a victim, or having been victimized at one point. I also don't really consider it to be healthy or productive to engage in "once a victim always a victim" mentality, because i think it sets a questionable precedent for mental health care. Semantically someone who was a victim once, is still a victim of that specific event, however it's a temporally relevant victimization, i just think people are getting a little loose on the usage of that word recently.
I'm still not sure how it makes that person a victim, unless it meets one of the described criteria i laid out, in which case it very explicitly becomes an abusive work. Otherwise it's questionable how you would even attribute victimization to the victim in question, because there is no explicit victim to even consider. I guess you could consider everybody even remotely tangentially relevant to be a victim, but that then opens a massive blackhole of logical reasoning which can't trivially be closed.
To propose a hypothetical here. Let's say there is a person who we will call bob. Bob has created a depiction of "abuse" in such a horrendous manner that even laying your eyes upon such a work will forever ruin you. We will define the work in question to be a piece of art, depicting no person in particular, arguably barely resembling a person at all, however the specific definition remains to the reader. You could hypothetically in this instance argue that even viewing the work is capable of making people a "victim" to it. However you want to work that one out.
The problem here, is that bob hasn't created this work in complete isolation, because he's just a person, he interacts with people, has a family, has friends, acquaintances, he's a normal person, aside from the horrors beyond human comprehension he has created. Therefore, in some capacity the influence of these people in his life, has to have influenced the work he engaged in on that piece. Are the people who know/knew bob, victims of this work as well, regardless of whether or not they have seen it, does the very act of being socially related to bob make them a victim of the work? For the purposes of the hypothetical we'll assume they haven't seen the work, and that he has only shown it to people he doesn't personally know.
I would argue, and i think most people would agree with me, that there is no explicit tie in between the work that bob has created, and the people he knows personally. Therefore, it would be a stretch to argue that because those people were tangentially relevant to bob, are now victims, even though they have not been influenced by it. Could it influence them in some external way, possibly causing some sort of external reaction? Yeah, that's a different story. We're not worried about that.
This is essentially the problem we have with AI, there is no explicit resemblance to any given person (unless defined, which i have already explicitly opted out of) or it has inherently based the image off of via training (which i have also somewhat, explicitly opted out of as well) there are two fundamental problems here that need to be answered. First of all, how are these people being victimized? By posting images publicly on the internet? Seems like they consented to people at least being aware of them, if not to some degree manipulating images of them, because there is nothing to stop that from happening (as everyone already knows from NFTs) And second of all, how are we defining these victims? What's the mechanism we use to determine the identity of these people, otherwise, we're just schizophrenically handwaving the term around calling people victims when we have no explicit way of determining that. You cannot begin to call someone a victim, if it's not even know whether they were victimized or not. You're setting an impossible precedent here.
Even if you can summarily answer those two questions in a decidedly explicit manner, it's still questionable whether that would even matter. Because now you would have to demonstrate some form of explicit victimization and damage resulting from that victimization. Otherwise you're just making the argument of "it's mine because i said so"
The law exists to protect the ruling class while not binding them, and to bind the working class without protecting them.
again if you're schizo, sure.
Yes.
on a loosely defined basis, yeah, in some capacity it uses the likeness of that person, but to what degree? How significantly? If the woman in the mona lisa picture was 4% some lady the artist saw three times a week due to their habits/routine would that make them suddenly entitled to some of that art piece in particular? What about the rest of it? You're running down an endless corridor of infinitely unfalsifiable, and falsifiable statements. There is no clear answer here.
Exactly. So, without consent, it shouldn’t be used. Periodt.
you need to explicitly define consent, and use, because without defining those, it's literally impossible to even begin determining the end position here.
I refuse to debate ideas on how to make ethical CSAM with you.
Go find a pedo to figure out how you want to try to make CSAM, and you can well akshully all tou want.
all of my arguments have explicitly removed any form of anything closely resembling CSAM to the point of being illegal under existing law, or at the very least, extremely questionable.
The only thing i haven't excluded is the potential to use an AI trained explicitly on humans, with no children being in the dataset, being used to generate porn of someone "under the age of 18" which it has zero basis of reality on, and cannot functionally do. That would be the only actual argument i can think of where that wouldn't already be illegal, or at least extremely questionable. Everything else i've provided a sufficient exclusion for.
Have fun calling me a pedo for no reason though, i guess.
Refresh for a new fake person
this ones a classic.
I've found that there's a lot of things on the Internet that went wrong because it was ad supported for "free". Porn is one of them.
There is ethically produced porn out there, but you're going to have to pay for it. Incidentally, it also tends to be better porn overall. The versions of the videos they put up on tube sites are usually cut down, and are only part of their complete library. Up through 2012 or so, the tube sites were mostly pirated content, but then they came to an agreement with the mainstream porn industry. Now it's mostly the studios putting up their own content (plus independent, verified creators), and anything pirated gets taken down fast.
Anyway, sites like Crash Pad Series, Erika Lust, Vanessa Cliff, and Adulttime (the most mainstream of this list) are worth a subscription fee.
Whats illegal in real porn should be illegal in AI porn, since eventually we won’t know whether it’s AI
That's the same as saying we shouldn't be able to make videos with murder in them because there is no way to tell if they're real or not.
That’s a good point, but there’s much less of a market for murder video industry
I mean, a lot of TV has murders in it. There is a huge market for showing realistic murder.
But I get the feeling your saying that there isn't a huge market for showing real people dying realistically without their permission. But that's more a technicality. The question is, is the content or the production of the content illegal. If it's not a real person, who is the victim of the crime.
Yeah the latter. Also murder in films for the most part is for storytelling. It’s not murder simulations for serial killers to get off to, you know what I mean?
without a victim
It was trained on something.
It can generate combinations of things that it is not trained on, so not necessarily a victim. But of course there might be something in there, I won't deny that.
However the act of generating something does not create a new victim unless there is someones likeness and it is shared? Or is there something ethical here, that I am missing?
(Yes, all current AI is basically collective piracy of everyones IP, but besides that)
Watching videos of rape doesn't create a new victim. But we consider it additional abuse of an existing victim.
So take that video and modify it a bit. Color correct or something. That's still abuse, right?
So the question is, at what point in modifying the video does it become not abuse? When you can't recognize the person? But I think simply blurring the face wouldn't suffice. So when?
That's the gray area. AI is trained on images of abuse (we know it's in there somewhere). So at what point can we say the modified images are okay because the abused person has been removed enough from the data?
I can't make that call. And because I can't make that call, I can't support the concept.
I mean, there’s another side to this.
Assume you have exacting control of training data. You give it consensual sexual play, including rough play, bdsm play, and cnc play. We are 100% certain the content is consensual in this hypothetical.
Is the output a grey area, even if it seems like real rape?
Now another hypothetical. A person closes their eyes and imagines raping someone. “Real” rape. Is that a grey area?
Let’s build on that. Let’s say this person is a talented artist, and they draw out their imagined rape scene, which we are 100% certain is a non-consensual scene imagined by the artist. Is this a grey area?
We can build on that further. What if they take the time to animate this scene? Is that a grey area?
When does the above cross into a problem? Is it the AI making something that seems like rape but is built on consensual content? The thought of a person imagining a real rape? The putting of that thought onto a still image? The animating?
Or is it none of them?
Is the output a grey area, even if it seems like real rape?
on a base semantic and mechanic level, no, not at all. They aren't real people, there aren't any victims involved, and there aren't any perpetrators. You might even be able to argue the opposite, that this is actually a net positive, because it prevents people from consuming real abuse.
Now another hypothetical. A person closes their eyes and imagines raping someone. “Real” rape. Is that a grey area?
until you can either publicly display yours, or someone else process of thought, or read peoples minds, definitionally, this is an impossible question to answer. So the default is no, because it's not possible to be based in any frame of reality.
Let’s build on that. Let’s say this person is a talented artist, and they draw out their imagined rape scene, which we are 100% certain is a non-consensual scene imagined by the artist. Is this a grey area?
assuming it depicts no real persons or identities, no, there is nothing necessarily wrong about this, in fact i would defer back to the first answer for this one.
We can build on that further. What if they take the time to animate this scene? Is that a grey area?
this is the same as the previous question, media format makes no difference, it's telling the same story.
When does the above cross into a problem?
most people would argue, and i think precedent would probably agree, that this would start to be a problem when explicit external influences are a part of the motivation, rather than an explicitly internally motivated process. There is necessarily a morality line that must be crossed to become a more negative thing, than it is a positive thing. The question is how to define that line in regards to AI.
We already allow simulated rape in tv and movies. AI simply allows a more graphical portrayal.
Consensual training data makes it ok. I think AI companies should be accountable for curating inputs.
Any art is ok as long as the artist consents. Even if they're drawing horrible things, it's just a drawing.
Now the real question is, should we include rapes of people who have died and have no family? Because then you can't even argue increased suffering of the victim.
But maybe this just gets solved by curation and the "don't be a dick" rule. Because the above sounds kinda dickish.
I see the issue with how much of a crime is enough for it to be okay, and the gray area. I can't make that call either, but I kinda disagree with the black and white conclusion. I don't need something to be perfectly ethical, few things are. I do however want to act in a ethical manner, and strive to be better.
Where do you draw the line? It sounds like you mean no AI can be used in any cases, unless all the material has been carefully vetted?
I highly doubt there isn't illegal content in most AI models of any size by big tech.
I am not sure where I draw the line, but I do want to use AI services, but not for porn though.
It just means I don't use AI to create porn. I figure that's as good as it gets.
It's not just AI that can create content like that though. 3d artists have been making victimless rape slop of your vidya waifu for well over a decade now.
Yeah, I'm ok with that.
AI doesn't create, it modifies. You might argue that humans are the same, but I think that'd be a dismal view of human creativity. But then we're getting weirdly philosophical.
With this logic, any output of any pic gen AI is abuse.. I mean, we can 100% be sure that there are CP in training data (it would be a very bug surprise if not) and all output is result of all training data as far as I understand the statistical behaviour of photo gen AI.
There is no ethical consumption while living a capitalist way of life.
ML always there to say irrelevant things
😆as if this has something to do with that
But to your argument: It is perfectly possible to tune capitalism using laws to get veerry social.
I mean every “actually existing communist country” is in its core still a capitalist system, or how you argue against that?
We could be sure of it if AI curated it's inputs, which really isn't too much to ask.
Well AI is by design not able to curate its training data, but companies training the models would in theory be able to. But it is not feasible to sanitise this huge stack of data.
With this logic, any output of any pic gen AI is abuse
Yes?
Watching videos of rape doesn’t create a new victim. But we consider it additional abuse of an existing victim.
is this a legal thing? I'm not familiar with the laws surrounding sexual abuse, on account of the fact that i don't frequently sexually abuse people, but if this is an established legal precedent that's definitely a good argument to use.
However, on a mechanical level. A recounting of an instance isn't necessarily a 1:1 retelling of that instance. A video of rape for example, isn't abuse anymore so than the act of rape within it, and of course the nonconsensual recording and sharing of it (because it's rape) distribution of that could necessarily be considered a crime of it's own, same with possession, however interacting with the video i'm not sure is necessarily abuse in it's own right, based on semantics. The video most certainly contains abuse, the watcher of the video may or may not like that, i'm not sure whether or that should influence that, because that's an external value. Something like "X person thought about raping Y person, and got off to it" would also be abuse under the same pretense at a certain point. There is certainly some interesting nuance here.
If i watch someone murder someone else, at what point do i become an accomplice to murder, rather than an additional victim in the chain. That's the sort of litmus test this is going to require.
That’s the gray area. AI is trained on images of abuse (we know it’s in there somewhere).
to be clear, this would be a statistically minimal amount of abuse, the vast majority of adult content is going to be legally produced and sanctioned, made public by the creators of those videos for the purposes of generating revenue. I guess the real question here, is what percent of X is still considered to be "original" enough to count as the same thing.
Like we're talking probably less than 1% of all public porn, but a significant margin, is non consensual (we will use this as the base) and the AI is trained on this set, to produce a minimally alike, or entirely distinct image from the feature set provided. So you could theoretically create a formula to determine how far removed you are from the original content in 1% of cases. I would imagine this is going to be a lot closer to 0 than it is to any significant number, unless you start including external factors, like intentionally deepfaking someone into it for example. That would be my primary concern.
That’s the gray area. AI is trained on images of abuse (we know it’s in there somewhere). So at what point can we say the modified images are okay because the abused person has been removed enough from the data?
another important concept here is human behavior as it's conceptually similar in concept to the AI in question, there are clear strict laws regarding most of these things in real life, but we aren't talking about real life. What if i had someone in my family, who got raped at some point in their life, and this has happened to several other members of my family, or friends of mine, and i decide to write a book, loosely based on the experiences of these individuals (this isn't necessarily going to be based on those instances for example, however it will most certainly be influenced by them)
There's a hugely complex hugely messy set of questions, and answers that need to be given about this. A lot of people are operating on a set of principles much too simple to be able to make any conclusive judgement about this sort of thing. Which is why this kind of discussion is ultimately important.
It can generate combinations of things that it is not trained on, so not necessarily a victim. But of course there might be something in there, I won’t deny that.
the downlow of it is quite simple, if the content is public, available for anybody to consume, and copyright permits it (i don't see why it shouldn't in most cases, although if you make porn for money, you probably hold exclusive rights to it, and you probably have a decent position to begin from, though a lengthy uphill battle nonetheless.) there's not really an argument against that. The biggest problem is identity theft and impersonation, more so than stealing work.
yeah bro wait until you discover where neural networks got that idea from
without a victim
You are wrong.
AI media models has to be trained on real media. The illegal content would mean illegal media and benefiting ,supporting, & profiting from and to victims of crime.
The lengths and fallacies pedophiles will go to justify themselves is absurd.
Excuse me? I am very offended by your insinuations here. It honestly makes me not want to share my thought and opinions at all. I am not in any way interested in this kind of content.
I encourage you to read my other posts in the different threads here and see. I am not an apologist, and do not condone it either.
I do genuinely believe AI can generate content it is not trained on, that's why I claimed it can generate illegal content without a victim. Because it can combine stuff from things it is trained on and end up with something original.
I am interested in learning and discussing the consequences of an emerging and novel technology on society. This is a part of that. Even if it is uncomfortable to discuss.
You made me wish I didn't..
Don't pay any attention to that kinda stupid comment. Anyone posting that kind of misinformation about AI is either trolling or incapable of understanding how generative AI works.
You are right it is a victimless crime (for the creation of content). I could create porn with minions without using real minion porn to put the randomnest example I could think of. There's the whole defamation thing of publishing content without someone's permission but that I feel is a discussion irrelevant of AI (we could already create nasty images of someone before AI, AI just makes it easier). But using such content for personal use... It is victimless. I have a hard time thinking against it. Would availability of AI created content with unethical themes allow people to get that out of their system without creating victims? Would that make the far riskier and horrible business of creating illegal content with real unwilful people disappear? Or at the very least much more uncommon? Or would make people more willing to consume thw content creating a feelibg of fake safety towards content previously illegal? There's a lot of implications that we should really be thinking about and how it would affect society, for better or worse...
Don't pay any attention to that kinda stupid comment. Anyone posting that kind of misinformation about AI is either trolling or incapable of understanding how generative AI works
Thank you 😊
