Technology

Oh I just assumed that every Conservative jerks off to kids

A 99-1 vote to drop the anti AI regulation is hardly the government voting against. The Senate smashed that shit hard and fast.

Even in countries a lot less corrupt than the US this is an issue.

Especially because the US government/companies doesn't do jack shit for people

In general, even up here in woke-ville, punishments have gotten a lot more strict for kids. There’s a lot more involvement of police, courts, jail. As a parent it causes me a lot of anxiety - whatever happened to school being a “sandbox” where a kid can make mistakes without adult consequences, without ruining their lives? Did that ever exist?

As a father of teenage girls, I don't necessarily disagree with this assessment, but I would personally see to it that anyone making sexual deepfakes of my daughters is equitably and thoroughly punished.

Punishment for an adult man doing this: Prison

Punishment for a 13 year old by doing this: Publish his browsing and search history in the school newsletter.

Schools generally means it involves underage individuals, which makes any content using them csam. So in effect, the "AI" companies are generating a ton of csam and nobody is doing anything about it.

Yes, finding out that your peers have been sharing deep fake pornography of you is absolutely fine and a normal thing for young girls to go through in school. No girls have ever killed themselves because of this exact sort of thing, surely. This definitely will not add in any way to the way women and girls are made to feel entirely disgustingly dehumanized by every man or boy in their lives. Groups of men and boys reducing them and their bodies down to vivid sexual fantasies that they can quickly generate photo realistic images of.

If the person in the image is underaged then it should be classified as child pornography. If the woman who's photo is being used hasnt consented to this then it should be classified as sexual exploitation.

Women and girls have faced degrees of this kind of sexual exploitation by men and boys since the latter half of the 20th century. But this is a severe escalation in that behavior. It should be illegal to do this and it should be prosecuted when and where it is found to occur.

I hope it might lead to a situation of dirty pics/vids not being a problem for the people in it any more, as it could be a deepfake. Like there were cases where a surfacing dirty pic was used for blackmail, ruined someones career or got them kicked out of some comittee, but since it could be fabrication now, I hope this will beva thing of the past, soon.

To add to that. I live in a red area and since the election I’ve been cat called much more. And it’s weird too, cus I’m middle aged…. I thought I’d finally disappear…

That's just called the outside now. Assume you are on camera at all times the moment you step out the front door. To be safe in the surveillance we live in today, best act as though you are being recorded in your own home as well.

Yeah there’s some nasty shit here. Big yikes, Lemmy.

I agree with the other comments, but wanted to add how deepfakes work to show how simple they are, and how much less information they need than LLMs.

Step 1: Basically you take a bunch of photos and videos of a specific person, and blur their faces out.

Step 2: This is the hardest step, but still totally feasable for a decent home computer. You train a neural network to un-blur all the faces for that person. Now you have a neural net that's really good at turning blurry faces into that particular person's face.

Step 3: Blur the faces in photos/videos of other people and apply your special neural network. It will turn all the blurry faces into the only face it knows how, often with shockingly realistic results.

not necessarily. image generation models work on a more fine-grained scale than that. they can seamlessly combine related concepts, like "photograph"+"person"+"small"+"pose" and generate plausible material due to the fact that all of those concepts have features in common.

you can also use small add-on models trained on very little data (tens to hundreds of images, as compared to millions to billions for a full model) to "steer" the output of a model towards a particular style.

you can make even a fully legal model output illegal data.

all that being said, the base dataset that most of the stable diffusion family of models started out with in 2021 is medical in nature so there could very well be bad shit in there. it's like 12 billion images so it's hard to check, and even back with stable diffusion 1.0 there was less than a single bit of data in the final model per image in the data.