Teen boys use AI to make fake nudes of classmates, sparking police probe
arstechnica.com
arstechnica.com
Bullies suck, and the kids who did this should get in some kind of trouble with the school and their parents, but this is not something police need to be involved in.
The moral panic over deepfakes is ridiculous, but the endless push to give literal children police records for acting like dumb kids is what's actually dangerous.
I never said it "should slide". This sort of behaviour is common in children and childhood is exactly the time to teach them to be better. This is a matter for the school and the parents to handle. There is zero reason for police to be involved in any of it.
It is literally bullying
> I suspect that at some point between the age of 13-19 as in this case there should be more serious repercussions.
Child development doesn't advance on a set schedule. Kids at the same age can be at vastly different levels of maturity and even older kids often make stupid choices carelessly. It's part of growing up. Their brains are a work in progress and that applies most to the parts that help them make good choices (https://www.aacap.org/AACAP/Families_and_Youth/Facts_for_Fam...).
That doesn't mean that kids can't make good choices, but it does mean we need to expect them to make terrible ones from time to time. As adults, it's our job to show them how they screwed up and set them right, but destroying their lives when their brains let them down isn't helping anyone.
Children need teaching and guidance as they grow up, not prison sentences or sex offender status. We gain nothing by treating children like criminals for acting like kids.
The moral panic might be ridiculous but taking some time to research the impacts so far should be a little sobering to a generalization of an interpreted generalization.
In this case, an extremely straw law might go a bit further than trying to suppress or control technology.
Eh, this is still a civil matter. Unless the picture was made for the purpose of harassment, there is no ill will we can presuppose the perpetrators had towards the victim. Just immense immaturity and a lack of judgement and empathy. Judgement and empathy can be taught. (If it was done with the intent of causing distress, i.e. cruelly, it's a matter for the police.)
I'd need to see more to conclude harassment, i.e. proof of intent to cause psychological harm, versus teasing without any clue about the impacts of one's actions.
> 'boys will be boys' mentality
I don't think this is equivalent. This is immature kids playing with new technology and pushing the boundaries of decency with it. Without proof of intended malice, or evidence of actual harm, it's not a matter for criminal procedure.
Teasing, whether you have a clue or not, is intent to cause psychological harm.
> Without proof of intended malice, or evidence of actual harm, it's not a matter for criminal procedure.
I concur. However I think it absolutely is grounds for a civil matter.
As someone who teased and was teased, was bullied and may have crossed into bullying, I disagree. The behaviours, when done with a group, are the seeking of social approval. The victim is closer to irrelevant than targeted. Where it crosses into bullying or harassment is where the victim is germane.
Yes, that's why invesitgations happen.
Source? It's a shower. Where are you hiding the phone?
I guess the FBI might knock down my door soon after trying to Google for some report or stats about it. Just some random news articles. Maybe I overestimated the problem?
Well, yes, that's clearly true whether or not it is also a criminal matter.
> Unless the picture was made for the purpose of harassment, there is no ill will we can presuppose the perpetrators had towards the victim
That is almost tautologically true, but on what bases do you conclude that it is so certain as not to warrant inquiry that there was no intent to harrass?
I conclude it is uncertain enough to warrant solely civil engagement on the basis of them being kids. Otherwise, every five year old pantsing their sibling becomes a matter for armed law enforcement.
And that's what laws should be concerned about. Creating something for your private consumption is usually a private matter. Distributing it should be judged according to the potential harm, as well as to potential legitimate uses.
I'll admit that showing your buddies a deepfake/photoshop/cut and paste job of that one kid you hate from social studies is very different from creating a website or fake social media profile where the whole world can see it. In this case, it seems like some pics were shared with classmates who deleted the photos when they were caught and now the police can't even find copies to see what all the fuss was about, but the internet adds an opportunity for harms that extend far beyond the classroom.
There might be some instances where a situation requires intervention beyond parents and teachers, but I see no evidence that this was one of them.
Dicey grounds here, but it should be legal to imagine _anything_. Thoughtcrimes are always wrong, and there never should be any law on what you can think or imagine.
However, even back in the day there has been quite the spectrum of just how bad this kind of thing is: one image of a crush, who is 17 (technically underage), shared with a best friend? Inappropriate but not across the line of "law enforcement action" even if the crush's parent is very offended.
A google drive link with 100s of pornographic deepfakes, of a single person (or sometimes many students) from a school, many of them 12-16 years old, shared extensively across the internet? This is a proper serious problem. Kinda like the difference between throwing a rock at a bus vs detonating a bomb in a train station. Both are some amount of kinetic force, but the magnitude is wildly different.
I'm not pretending that physical violence is the same as making pornographic imagery, but my experience in that field (no jokes plz!) shows a much stronger link than many would like to admit, from one to the next. It's useful for teenagers to learn where these boundaries are in their particular society - life is often more pleasant when ~everyone agrees on the local standards of behaviour and they are (mostly) followed.
As for police action where I lived, it too followed a broad spectrum: on one end, there is arrest and court action which on occasion is definitely warranted (another pun, sorry). On the flipside, there's educating kids about the reality of the world they live in: If you share a nude, it might get shared again -- consider not including your face and your reproductive organs in the same picture. If someone generates a fake picture of you, it'll feel hurtful and you will be wronged, but what if they wrote something mean (and fake) about you? Eventually you will need to deal with the fact that some people are arses.
Some combination of protecting victims (but also helping them grow a thick skin in this world), educating kids on what is appropriate so they can know there is some boundary where things move from 'funny' to 'wrong'. They usually already have a sense of 'private' vs 'public', though that distinction is a becoming lot blurrier/different for current young generations.
And nothing above has anything really to do with 'AI'.
Rather than burying our heads in the sand and making it illegal to produce those, I think the better long term approach is to teach our children to be skeptical, and to emphasize that we can't believe everything that's shared on the internet.
On the plus side, the advancing technology can provide plausible deniability if a genuine intimate photo/video gets leaked, although as long as we're hoping for humanity to evolve a little, maybe we can collectively get a little less prudish about the human body and sex too.
Yup. We're already sort of getting there.
I'm already living in a world where I can't trust what I read. Soon I won't be able to trust what I see or hear online. Face to face interactions will be required for anything important.
Not looking forward to this.
I hope we get to a place where the automatic assumption upon seeing any digital image is that it's probably fake.
If these kids had raped and murdered I'd agree that that's an issue teachers and parents need outside help to address, but that's not what we're talking about.
> These are also teens rather than children.
Teens are adolescent children and their undeveloped brains make them highly susceptible to making poor choices and failing to consider the consequences of their actions. This is the time of their lives when they're naturally going to do things that they shouldn't.
There's a reason we don't lock schoolyard bullies up for theft when they steal someone's lunch money. We don't arrest them for doxing when they write a classmates name and phone number next to "for a good time call" on the bathroom stall. That's stupid kid shit that children have been doing for ages and the solution isn't prison, it's parenting, so that those kids can learn from their poor choices and become better people.
Teens can be held accountable for their actions without involving law enforcement.
Neither can fraud or murder. Yet we still ban it. This concept that if a problem cannot be 100% solved by the law, it shouldn't be banned, is an umbrella for the insufferable.
You can make a photorealistic drawing too, that can pass as a real photography (and with enough skill it will be better than whatever generative models can produce). The difference here is the lower barrier to make the images.
But I think that even a photorealistic drawing, if made and distributed to others, should be considered a form of harassment
I agree, but critically, I think it's the harassment that matters, not the drawing.
If someone draws something in the privacy of their own home (maybe or maybe not with the assistance of an AI), and the police find it in an unrelated search, they should have nothing to prosecute. Drawing is one way humans work through their thoughts and emotions; the act of putting pencil to paper—including digital pencil to digital paper—should never be illegal, lest we effectively create thought crime.
I'm not trying to say it's ok, but just like the person you replied to, I truly find these things comparable.
I don‘t see the argument here. What you describe is sexual harassment and should be illegal.
If person A draws a naked picture and shares it with person B, and it looks like person C, then person C has been sexually harassed?
If you share it with the entire class? In the school building itself? It becomes impossible for the target to not be aware.
The moment someone teases the other it doesn’t matter how many times it’s been shared.
If the person is recognizable, then in many cases it would violate existing nonconsensual/revenge porn laws, which very often are not concerned with provenance but whether its a explicit, a recognizable depiction, and nonconsensual.
(Depending on where it is shared it might, as others note, be sexual harassment as well.)
You can develop fake images of someone from pretty much purely public data. Images in the past, you would have needed to take an actual picture of the person. And if you had taken a picture of the person, there's generally a way to handle this scenario under copyright laws.
If someone shows up in court and says it's not a picture of the person suing, and just a generated image of someone who looks rather similar (people looking similar to each other being a rather common occurrence), and that the image is unlikely to cause imminent lawless action, it would be really interesting to watch the court try to pick apart what's going on.
So the AI is going to take what you give it, and combine that with what others have given it based on a bunch of similarity weights. If the AI is any good at its job, it will produce an image that is remarkably correlated to the undisclosed original, since that is exactly what AI is doing every time it's asked to do anything at all: "I have an incomplete thought/image in my mind, please produce a written or visual completion of it based on a model built on several billion best guesses from real world data." And if the victim of the harassment has had the misfortune (or lack of foresight) to have their actual nude images somehow slip into the ai's training data, it is likely the AI will prioritize surfacing them in its output.
Either way, the AI will now give you a body that is roughly correlated with what your head looks like.
I would assume the girls - as classmates - are of a similar are to the "teen boys". There's unlikely to be a significant age differential, and it doesn't really change the ethical implications.
> While some legal authorities argue against the legality of these AI-generated depictions, there is no established precedent for prosecuting their creation.
It's possible that what they did was illegal regardless of the age of the girls. These laws however exist to protect children. Treating a bunch of hormonal teenagers the same as a 50 year old adult is unlikely to be a net positive to society.