Kinda weird that it details how badly this affected the girls' mothers. The girls don't get a say, but won't someone please think of the mothers?!
It is pretty weird, like "The reaction was one of massive support for all the affected mothers"..?
How do the girls not get a say? They asked their mothers for help who organised to found others who are affected.
I imagine it leans into the idea of some people being "too young" to form a grown-up opinion.
Really fucking weird, given the context is around their likeness being used for the purpose of porn.
Maybe be We should shift our thinking to assume that everything posted on the internet is Fake. that's the only solution to counter the proliferation of AI. the genie is out of the bottle and can't be forced back
only believe information from official sources that is cryptographically signed.
This is a a really really over simplified solution and I’m gonna argue it’s not at all effective. The cat is out of the bag, just like you said. You can’t undo that.
Nothing on the internet is real, okay let’s start from there.
So now how do we relay scientific findings for example? Rely on the media? Pray to some god and hope our reasoning and interpretations are correct enough?
Should we trust video recordings? Pictures?
Should we trust word of mouth? Each other?
Ourselves?
Eh, most dashcams have metadata with GPS, timestamps, etc.
GPS locations, time, that's just math tho. But we could put a private key on every camera and digitally sign every photo/video to prove where it came from.
If it gets bad enough you'll just carry around a film camera and snap a photo of an accident to prove your dashcam was "real."
Official news sources, such as the AP, aren't going to just start faking shit. So look towards media that has a reputation. Yellow journalism has always been around and always will. But one advacent in faking whatever does not mean countermeasures stop advancing.
But we could put a private key on every camera and digitally sign every photo/video to prove where it came from.
Unfortunately, the movie Freeze Frame is becoming increasingly relevant.
I’m going to push back mainly on your last point. “News agencies are just going to start faking shit”
Why not? What happened to CNN? What’s to stop anyone from buying out the news agency and steering the ship elsewhere? Obviously they wouldn’t want to be obvious about it but a little change here and there and you’d be surprised what you can convince people of.
As a society we are incredibly susceptible to social influence and it’s only going to get easier. I can’t rely on the past to predict the future accurately when the medium is so easily changeable.
that's where things are headed anyways. AI is only getting better, and very soon every one will be suing everyone else for label and stuff.
OpenAI have recognized that they don't have the means to distinguish between human and AI generated essays. and soon pics and videos generated will be just as hard to verify. I say we should only trust what we can have the means to verify or go back to old tech, like public broadcasting channels for official news and announcements.
(Didn’t realize we had simultaneous convos going lmao hi again)
Can you clarify a bit on this one? What’s headed where specifically? What’s your claim?
Again though, like you said the cat is out of the bag. You can slap all the laws you want on AI. The technology and the concept is already created and been fleshed out. You can’t destroy information and concepts. So to ban it or somehow convince everyone to abandon it is doomed to fail immediately and intrinsically. The powerful people in the world (like my other comment) will still be able to utilize this very powerful technology. Which puts everyone else at a significant and arguably irreparable disadvantage.
I think AI is no different from any other emerging technology, look at nukes! With this great power comes great responsibility that we need to flesh out and figure out.
For example, AI generate image or not, deep fake nudes are still sexual harassment (not physical of course but no different from revenge porn) they take away someone’s autonomy and thus we treat the offenders no differently. Yes the threat is there but we acknowledge it and we have a way to punish for it.
Less trying to make traps to send the cat back into the bag and more hunters and techniques to catch all the kittens that the cat will create. We are very intelligent beings all things considered, we can handle this stuff we just have to learn from our past and make meaningful changes the best we can and we will continue to thrive as we always have. (Assuming we can fix climate change or prepare accordingly and adequately)
Thanks to deepfakes basically none of it is trustworthy anymore.
Violates the right to your own image. You are not allowed to upload images of a classmate to an AI cloud without asking and neither to reach the generated images around.
Is that an actual legal right? If you've described it accurately, then Facebook and Instagram would be completely illegal
It depends on your location, different countries have very different laws.
For example in most countries it's perfectly acceptable to have someone in a picture that you're taking in public (for example you're taking a picture of a building and someone happens to walk by). A notable exception to this is France, where apparently the right to ones own image is quite strong which technically makes most pictures of the Eiffel Tower illegal (as long as any one person is identifiable on it).
Taking (and distributing) a picture specifically of a specific person that's just doing random stuff in public is already less uniform and varies. There's often some protection to basically say "no, you can't make fun of some random person for having the wrong tshirt, they have a right to privacy". A notable exception to that is usually "public figures" (which mostly means people in political, religious or commercial leadership positions): they mostly just have to accept to be pictured wherever.
Protection for pictures taken in a private is usually the strongest (so yes, if you post a picture of your 3 best friends at a small party in your home, you might have to ask them for permission!)
How all of this applies to pictures that "aren't real" but look disturbingly so is probably going to be fought over in court for a good while.
No, human right. And yeah, they mostly are. But it's not Facebook offending but each of the teens, so nobody can really enforce it. Same like with phone numbers, except that those are actually protected by law in most countries.
AI generates novel images, though. They are merely trained to produce your likeness. None of the pixels are from any source images.
In this case, I'm mistaken. They used a clothing remover app on normal photos and did not train an AI.
Jesus Christ that’s not even close to AI they literally stitched together shit like photoshop
By the way, where do you see that clothoff ( the app mentioned in the article) doesn’t use trained AI models? I’m refraining from visiting their site to check myself as I don’t really want to give them that traffic and I figured I’d ask you direct instead as you already went their to verify I assume)
I didn't say the app doesn't use trained models. I said the students didn't themselves train an embedding or LORA against the other students' faces in order to generate entirely new pics.
Ah, good point I missed that detail. Thanks for clarifying
You can't stop them being made, they're just the same deepfakes people have been making before. It's important to note that they're not photos of people, they're guesses made by a algorithm.
While you're completely right, that's hardly a consolation for those affected. The damage is done, even if it's not actually real, because it will be convincing enough for at least some.
While I understand your point, what consolation can be provided?
I think the people who made the pictures have to suffer consequences. Otherwise this sends the message as if it was just fair game to behave that way.
Saying man that sucks doe, and nothing else really
The faces are not generated, and that is where the damage comes. It targets the girls for humiliation by implying that they allowed the nudes to be taken of them. Depending upon the location and circumstances, this could get the girls murdered. Think of "honor killings" by fundamentalists. It makes them targets for further sexual abuse, too.
Anyone distributing the photos are at fault, as well as the people who made the photos.
The problem goes deeper, though. We can never trust a photo as proof of anything, again. Let that sink in, what it means to society.
To push back your attempt to minimalise what's going on here ...
Yes, they're not actually photos of the girls. But, nor is a photo of a naked person actually the same as that person standing in front of you naked.
If being seen naked is unwanted and embarrassing etc, why should a photo of you naked be embarrassing, and, to make my point, what difference would it make if the photo is more or less realistic? An actual photo can be processed or taken under certain lighting or with a certain lens or have been taken some time in the past ... all factors that lessen how close it is to the current naked appearance of the subject. How unrealistic can a photo be before it's no longer embarrassing?
Psychologically, I'd say it's pretty obvious that the embarrassment of a naked image is that someone else now has a relatively concrete image in their minds of what the subject looks like naked. It is a way of being seen naked by proxy. A drawn or painted image could probably have the same effect.
There's probably some range of realism within which there's an embarrassing effect, and I'd bet AI is very capable of getting in that range pretty easily these days.
While the technology is out there now ... it doesn't mean that our behaviours with it are automatically acceptable. Society adapts to the uses and abuses new technology has and it seems pretty obvious that we're yet to culturally curb the abuses of this technology.
Tldr
Exactly, the technology is out there and will not cease to exist. Maybe we'll digitally sign our photos in the future so that deepfakes can be sorted out by that.
Omg it's NFTs time to shine!!!!
/S
Will everyone be expected to have some kind of official PGP key?
Kinda weird that it details how badly this affected the girls' mothers. The girls don't get a say, but won't someone please think of the mothers?!
It is pretty weird, like "The reaction was one of massive support for all the affected mothers"..?
How do the girls not get a say? They asked their mothers for help who organised to found others who are affected.
I imagine it leans into the idea of some people being "too young" to form a grown-up opinion.
Really fucking weird, given the context is around their likeness being used for the purpose of porn.
Maybe be We should shift our thinking to assume that everything posted on the internet is Fake. that's the only solution to counter the proliferation of AI. the genie is out of the bottle and can't be forced back
only believe information from official sources that is cryptographically signed.
This is a a really really over simplified solution and I’m gonna argue it’s not at all effective. The cat is out of the bag, just like you said. You can’t undo that.
Nothing on the internet is real, okay let’s start from there.
So now how do we relay scientific findings for example? Rely on the media? Pray to some god and hope our reasoning and interpretations are correct enough?
Should we trust video recordings? Pictures?
Should we trust word of mouth? Each other? Ourselves?
Eh, most dashcams have metadata with GPS, timestamps, etc.
GPS locations, time, that's just math tho. But we could put a private key on every camera and digitally sign every photo/video to prove where it came from.
If it gets bad enough you'll just carry around a film camera and snap a photo of an accident to prove your dashcam was "real."
Official news sources, such as the AP, aren't going to just start faking shit. So look towards media that has a reputation. Yellow journalism has always been around and always will. But one advacent in faking whatever does not mean countermeasures stop advancing.
Unfortunately, the movie Freeze Frame is becoming increasingly relevant.
I’m going to push back mainly on your last point. “News agencies are just going to start faking shit”
Why not? What happened to CNN? What’s to stop anyone from buying out the news agency and steering the ship elsewhere? Obviously they wouldn’t want to be obvious about it but a little change here and there and you’d be surprised what you can convince people of.
As a society we are incredibly susceptible to social influence and it’s only going to get easier. I can’t rely on the past to predict the future accurately when the medium is so easily changeable.
that's where things are headed anyways. AI is only getting better, and very soon every one will be suing everyone else for label and stuff.
OpenAI have recognized that they don't have the means to distinguish between human and AI generated essays. and soon pics and videos generated will be just as hard to verify. I say we should only trust what we can have the means to verify or go back to old tech, like public broadcasting channels for official news and announcements.
(Didn’t realize we had simultaneous convos going lmao hi again)
Can you clarify a bit on this one? What’s headed where specifically? What’s your claim?
Again though, like you said the cat is out of the bag. You can slap all the laws you want on AI. The technology and the concept is already created and been fleshed out. You can’t destroy information and concepts. So to ban it or somehow convince everyone to abandon it is doomed to fail immediately and intrinsically. The powerful people in the world (like my other comment) will still be able to utilize this very powerful technology. Which puts everyone else at a significant and arguably irreparable disadvantage.
I think AI is no different from any other emerging technology, look at nukes! With this great power comes great responsibility that we need to flesh out and figure out.
For example, AI generate image or not, deep fake nudes are still sexual harassment (not physical of course but no different from revenge porn) they take away someone’s autonomy and thus we treat the offenders no differently. Yes the threat is there but we acknowledge it and we have a way to punish for it.
Less trying to make traps to send the cat back into the bag and more hunters and techniques to catch all the kittens that the cat will create. We are very intelligent beings all things considered, we can handle this stuff we just have to learn from our past and make meaningful changes the best we can and we will continue to thrive as we always have. (Assuming we can fix climate change or prepare accordingly and adequately)
Thanks to deepfakes basically none of it is trustworthy anymore.
Violates the right to your own image. You are not allowed to upload images of a classmate to an AI cloud without asking and neither to reach the generated images around.
Is that an actual legal right? If you've described it accurately, then Facebook and Instagram would be completely illegal
It depends on your location, different countries have very different laws.
For example in most countries it's perfectly acceptable to have someone in a picture that you're taking in public (for example you're taking a picture of a building and someone happens to walk by). A notable exception to this is France, where apparently the right to ones own image is quite strong which technically makes most pictures of the Eiffel Tower illegal (as long as any one person is identifiable on it).
Taking (and distributing) a picture specifically of a specific person that's just doing random stuff in public is already less uniform and varies. There's often some protection to basically say "no, you can't make fun of some random person for having the wrong tshirt, they have a right to privacy". A notable exception to that is usually "public figures" (which mostly means people in political, religious or commercial leadership positions): they mostly just have to accept to be pictured wherever.
Protection for pictures taken in a private is usually the strongest (so yes, if you post a picture of your 3 best friends at a small party in your home, you might have to ask them for permission!)
How all of this applies to pictures that "aren't real" but look disturbingly so is probably going to be fought over in court for a good while.
No, human right. And yeah, they mostly are. But it's not Facebook offending but each of the teens, so nobody can really enforce it. Same like with phone numbers, except that those are actually protected by law in most countries.
AI generates novel images, though. They are merely trained to produce your likeness. None of the pixels are from any source images.In this case, I'm mistaken. They used a clothing remover app on normal photos and did not train an AI.
Jesus Christ that’s not even close to AI they literally stitched together shit like photoshop
By the way, where do you see that clothoff ( the app mentioned in the article) doesn’t use trained AI models? I’m refraining from visiting their site to check myself as I don’t really want to give them that traffic and I figured I’d ask you direct instead as you already went their to verify I assume)
I didn't say the app doesn't use trained models. I said the students didn't themselves train an embedding or LORA against the other students' faces in order to generate entirely new pics.
Ah, good point I missed that detail. Thanks for clarifying
You can't stop them being made, they're just the same deepfakes people have been making before. It's important to note that they're not photos of people, they're guesses made by a algorithm.
While you're completely right, that's hardly a consolation for those affected. The damage is done, even if it's not actually real, because it will be convincing enough for at least some.
While I understand your point, what consolation can be provided?
I think the people who made the pictures have to suffer consequences. Otherwise this sends the message as if it was just fair game to behave that way.
Saying man that sucks doe, and nothing else really
The EU is assessing the state of deepfakes and is planning to create a legal framework to keep it under control through a "combination of measures will likely be necessary to limit the risks of deepfakes, while harnessing their potential."
The faces are not generated, and that is where the damage comes. It targets the girls for humiliation by implying that they allowed the nudes to be taken of them. Depending upon the location and circumstances, this could get the girls murdered. Think of "honor killings" by fundamentalists. It makes them targets for further sexual abuse, too. Anyone distributing the photos are at fault, as well as the people who made the photos.
The problem goes deeper, though. We can never trust a photo as proof of anything, again. Let that sink in, what it means to society.
To push back your attempt to minimalise what's going on here ...
Yes, they're not actually photos of the girls. But, nor is a photo of a naked person actually the same as that person standing in front of you naked.
If being seen naked is unwanted and embarrassing etc, why should a photo of you naked be embarrassing, and, to make my point, what difference would it make if the photo is more or less realistic? An actual photo can be processed or taken under certain lighting or with a certain lens or have been taken some time in the past ... all factors that lessen how close it is to the current naked appearance of the subject. How unrealistic can a photo be before it's no longer embarrassing?
Psychologically, I'd say it's pretty obvious that the embarrassment of a naked image is that someone else now has a relatively concrete image in their minds of what the subject looks like naked. It is a way of being seen naked by proxy. A drawn or painted image could probably have the same effect.
There's probably some range of realism within which there's an embarrassing effect, and I'd bet AI is very capable of getting in that range pretty easily these days.
While the technology is out there now ... it doesn't mean that our behaviours with it are automatically acceptable. Society adapts to the uses and abuses new technology has and it seems pretty obvious that we're yet to culturally curb the abuses of this technology.
Tldr
Exactly, the technology is out there and will not cease to exist. Maybe we'll digitally sign our photos in the future so that deepfakes can be sorted out by that.
Omg it's NFTs time to shine!!!!
/S
Will everyone be expected to have some kind of official PGP key?
That is very messed up.
Inevitable. Our technology outpaced our evolution a long time ago. We’re spiraling.
I think that happened at least 10k years ago ... it's just that the spiral is getting faster and faster ...
That's always the case. Evolution, both biological and societal, happens after the environment changes.