An Investigation Found Hundreds Of Known Images Of Child Sexual Abuse Material (CSAM) In An Open Dataset

An investigation found hundreds of known images of child sexual abuse material (CSAM) in an open dataset used to train popular AI image generation models, such as Stable Diffusion. Models trained on this dataset, known as LAION-5B, are being used to create photorealistic AI-generated nude images, including CSAM. It is challenging to clean or stop the distribution of publicly distributed datasets as it has been widely disseminated. Future datasets could use freely available detection tools to prevent the collection of known CSAM. (SOURCE: https:// cyber.fsi.stanford.edu/news/investigation-finds-ai-image-generation-models-trained-child-abuse)

This isn't some "gotcha". This is something that actually HAPPENED.

Firstly;

Secondly, pick a fucking argument, dude. You've gone from asking me if I think AI rape images are fine to asking if I think AI CSA is fine because its using real CSEM to calling out that some AI datasets were contaminated.

So what's your argument? Are you arguing against my moral values or against AI? Those specific AI programs? People using art to fulfil unorthodox fantasy?

Copy-pasting statements after a furious little google search because I'm not bending over ass-backwards to agree with you is presenting what point, exactly?

AI bad?

Taboo kink bad?

Fictional content bad?

RPF bad?

Which is it? Help me out here. Is it some hybrid of all of those points?

Because I mean if your point is that AI is bad, I agree! Whole heartedly. For a multitude of reasons, not least one of your copied points; that it is actually basically impossible to make the AI forget something its learned. Hence; scrapping it all and starting over. Hence; stricter laws that only allow AI to be trained on specific open source datasets. Hence; requiring human approval for reference images and constant human monitoring rather than simply allowing the AI to run unsupervised, as most are.

I just don't agree that using various mediums to create fictional art that isn't created with a malicious directive is bad. Whether that's done by hand or using (god forbid) AI.

Last I checked tumblr doesn't allow links to be sent via ask, and this could be found with a Google search, but here. Hope this goes through. dontreblogbigots . tumblr . com/post/662890014546460672/prismatic-bell-is-a-zionist-among-other-things

Tumblr removed the ability to hyperlink in asks, but you can still just paste the link as plain text in the askbox. No need for spaces or breaks. At one point it had also removed the ability to paste a link, but that only lasted about three days.

I'll do my research on the matter. (The blog, not the linking.)

OP is also getting raked over hot coals for being a racist. I wouldn't worry about it.

I am actually begging some people to just let some spaces exist untouched by real-world issues and horrors.

Like I've lost count of the amount of times peaceful game or fandom servers have been ruined by people stampeding in with political rants, bitching about world issues, demanding internal activism, demanding vent channels so they can whine about their shitty parents, ect.

Like. Respectfully. Not every single space has to be inclusive of and welcoming of outside topics. The real world sucks. We don't needed to be reminded of that absolutely everywhere.

I am actually begging some people to just let some spaces exist untouched by real-world issues and horrors.

Like I've lost count of the amount of times peaceful game or fandom servers have been ruined by people stampeding in with political rants, bitching about world issues, demanding internal activism, demanding vent channels so they can whine about their shitty parents, ect.

Like. Respectfully. Not every single space has to be inclusive of and welcoming of outside topics. The real world sucks. We don't needed to be reminded of that absolutely everywhere.

What even is proshipping

Proshipping is pro-shipping. As in; for shipping. In support of shipping.

'Proship' is the modern evolution of the original acronym SALS, which stands for Ship And Let Ship. It is exactly the same concept as; "your kink is not my kink, but that's okay. Its simply not my business."

The core values of proshipping are:

Everyone has a right to ship.

Ships cannot and should not be controlled or dictated. (E.g; "I don't like this ship so you're a bad person for liking it and you're not allowed to ship it.")

People should not be shamed, harassed, threatened, ect because of their ships.

Modern 'proshipping' will often vary from individual to individual. In truth, the term 'proshipping' as its used and defined these days would be actually be 'profiction.' This is because many people expand proshipping in order encompass being pro-X across an array of varying creative and fandom based criteria.

However as 'proship' is the more commonly known label it is the most widely accepted.

There is often the common misconception stemming from careful propaganda that proshipping exclusively refers to the support and enjoyment of taboo or dark themes such as incest and pedophilia. This is not true.

Proshipping does not mean: "I enjoy incest." Proshipping means: "even if I do not enjoy incest, it is not my place to try to stop you doing so or harming you because of it."

Okay, fine, AI rape art is fine because it's not real. What about AI CSAM? Is it fine because it's art, or is it bad because real CSAM content is basically guaranteed to have been involved in training the AI? Again, where do you think AI /learns/ to make photorealistic RAPE art? It uses REFERENCES.

I can't believe you're making me somewhat defend AI generation here but off we go I guess.

AI data-mines accessible media and specific locations for media. While AI continuously learns the more you use it, the majority of AI's founding knowledge came pre-launch, as well as with its training. AI is not digging its little fingers around in the dark web.

For example if you wanted to prompt 'child porn' the AI does not immediately start digging around in the dark web for child pornography. It pairs the two keywords 'child' with everything its learned on children and 'porn' with everything its learned on sex and pornography and mashes them together.

So, no. Its not "basically guaranteed" to have used real CSEM. Nice try, though.

Granted there are still AI services which allow you to insert a sample image, but you're kidding yourself if you don't think an inserted reference image paired with specific prompts won't get flagged for review. AI techbros don't want their million dollar investments getting tanked and while they don't care about copyright you can bet they'll care if the FBI start kicking down their doors.

Most AI generators, in their legal statements, will have a section about using unlawful content to reference, train or manipulate their technology.

A lot of AI generation tools also blacklist certain key terms anyway so pornographic or "unsavory" content cannot be created. (Like rape images, murder images, gore images, ect.)

(E.g; if I tried to prompt 'Trump getting murdered' on select AI generation sites it would tell me I cannot use a prompt including 'murdered.')

I will say that there needs to be a better system in place for that. I'm sure a lot of samples of actual illegal activity have made it through the cracks (as it does everywhere), which is part of why I firmly believe AI generation should be shut down until the entire system can be reworked in order to offer better protection to the people who are being used as cashcows. In terms of copyright protection and more effective systems for preventing AI being trained off of certain content.

And because I know you're going to ask; yes. I do believe that if an image was generated sourced from actual CSEM, that the image should be as thoroughly scrubbed as possible. I do not condone CSEM in any form nor for any uses.

Nor do I condone AI generation (as it currently stands.)

Are you done with your little gotcha attempts now, Anon? I'm bored.