doesn’t it follow that AI-generated CSAM can only be generated if the AI has been trained on CSAM?
Not quite, since the whole thing with image generators is that they’re able to combine different concepts to create new images. That’s why DALL-E 2 was able to create a images of an astronaut riding a horse on the moon, even though it never saw such images, and probably never even saw astronauts and horses in the same image. So in theory these models can combine the concept of porn and children even if they never actually saw any CSAM during training, though I’m not gonna thoroughly test this possibility myself.
Still, as the article says, since Stable Diffusion is publicly available someone can train it on CSAM images on their own computer specifically to make the model better at generating them. Based on my limited understanding of the litigations that Stability AI is currently dealing with (1, 2), whether they can be sued for how users employ their models will depend on how exactly these cases play out, and if the plaintiffs do win, whether their arguments can be applied outside of copyright law to include harmful content generated with SD.
My question is: why aren’t OpenAI, Google, Microsoft, Anthropic… sued for possession of CSAM? It’s clearly in their training datasets.
Well they don’t own the LAION dataset, which is what their image generators are trained on. And to sue either LAION or the companies that use their datasets you’d probably have to clear a very high bar of proving that they have CSAM images downloaded, know that they are there and have not removed them. It’s similar to how social media companies can’t be held liable for users posting CSAM to their website if they can show that they’re actually trying to remove these images. Some things will slip through the cracks, but if you show that you’re actually trying to deal with the problem you won’t get sued.
LAION actually doesn’t even provide the images themselves, only linking to images on the internet, and they do a lot of screening to remove potentially illegal content. As they mention in this article there was a report showing that 3,226 suspected CSAM images were linked in the dataset, of which 1,008 were confirmed by the Canadian Centre for Child Protection to be known instances of CSAM, and others were potential matching images based on further analyses by the authors of the report. As they point out there are valid arguments to be made that this 3.2K number can either be an overestimation or an underestimation of the true number of CSAM images in the dataset.
The question then is if any image generators were trained on these CSAM images before they were taken down from the internet, or if there is unidentified CSAM in the datasets that these models are being trained on. The truth is that we’ll likely never know for sure unless the aforementioned trials reveal some email where someone at Stability AI admitted that they didn’t filter potentially unsafe images, knew about CSAM in the data and refused to remove it, though for obvious reasons that’s unlikely to happen. Still, since the LAION dataset has billions of images, even if they are as thorough as possible in filtering CSAM chances are that at least something slipped through the cracks, so I wouldn’t bet my money on them actually being able to infallibly remove 100% of CSAM. Whether some of these AI models were trained on these images then depends on how they filtered potentially harmful content, or if they filtered adult content in general.
I didn’t know there was a difference between mice and rats but after looking it up (read: asking ChatGPT), yes it was a small mouse. Also I forgot the air fryer half open at night.
Air fryers aren’t microwaves heating things in weird ways. It’s just an oven that gets hot faster and blows air around a lot. Why would a mouse explode?
idk man maybe the air inside the mouse expanded until it exploded or something? I can’t really give you a definitive answer since I didn’t dissect the little guy. All I know is that I heard something pop, and when I looked inside my air fryer there was a twitching dead mouse with a splash of blood coming out of its mouth.
I heard that he’s an ethereal being from another dimension that has already faded away from our plane of existence so the police is wasting its time looking for him and should close the case.
I second the Nerts recommendation because when I was a teen I played Ligretto (Basically just a branded Nerts) at one party and we all had a blast. To this day I wish I had the opportunity to play it again, but I’m not a party guy nor do I have that many decks of cards so /shrug
It’s a shame that it doesn’t seem to have that many players anymore. It’s a lot like other multi-class shooter games like Overwatch and Paladins, but about three times more crazy and weird, and I had a ton of fun playing it as a teen.
I’ve heard Garden Warfare 2 is pretty good too though, so I’ll consider buying it at some point (when I have money).
Not sure how deeply positive it was necessarily, but as far as I can tell I changed at least one person’s life significantly.
When I was in middle school I learned how to solve the Rubik’s cube through online tutorials, and afterwards I was basically addicted to it for a while and I was messing with a cube all the time. One of my classmates was interested in learning how to solve one, so I drew a bunch of guides and taught him the terminology, and soon enough he was also hooked on the Rubik’s cube.
I eventually grew tired of it, but he kept going and learned how to solve all sorts of smaller and bigger cubes and pyramids. I bumped into him at uni a few years later and he had a cube in his hand lol. Last I heard of him he was even participating in some local competitions, but idk what he’s up to now.
Even if he ultimately gave up on his speedcubing dreams I can at least say that I introduced him to a cool new hobby that kept him entertained for several years.
Winter because I live in Brazil, which is a tropical and very humid country so the heat can get unbearable even here in the south where it’s colder compared to the north.
There are a few cities that get cold enough to reach negative Cº, but where I live it’s mostly between 10º-20º during winter so I just need some extra coats and more blankets and I’m happy. During summer I can only take off so many layers before it becomes illegal.
Buy Minecraft, which apparently is on sale for $14.99. It’s not quite as fun when all your friends have the official version and get to play together while you’re stuck with an - *ahem* - unofficial version.
One that I really like is Dear David by Adam Ellis. And yeah yeah, I know that he’s going for the usual “true ghost story” marketing shtick or whatever, but these stories that don’t just use cheap jumpscares and instead build a sense of constant threat and impending doom are the ones that make it harder for me to sleep at night.
Capybaras are pretty common in the area where I live, and really throughout most of Brazil. Don’t get me wrong, we still think they’re pretty cute, but I’ve seen some Americans get really excited about them.
Oh, and the maned wolf. To be fair, I think they’re pretty neat too.
Although I don’t have kids everything I’ve heard about the Ruby Franke/8 Passengers case made my blood boil. That scumbag saw her kids as nothing but a way to get clicks and money, and I’m glad that she’s going to rot in prison.
And related to this, add animal abuse to the list of things that make my blood boil too. It’s wild to think that until recently there was such a thing as a thriving “monkey torture network”.
Not quite, since the whole thing with image generators is that they’re able to combine different concepts to create new images. That’s why DALL-E 2 was able to create a images of an astronaut riding a horse on the moon, even though it never saw such images, and probably never even saw astronauts and horses in the same image. So in theory these models can combine the concept of porn and children even if they never actually saw any CSAM during training, though I’m not gonna thoroughly test this possibility myself.
Still, as the article says, since Stable Diffusion is publicly available someone can train it on CSAM images on their own computer specifically to make the model better at generating them. Based on my limited understanding of the litigations that Stability AI is currently dealing with (1, 2), whether they can be sued for how users employ their models will depend on how exactly these cases play out, and if the plaintiffs do win, whether their arguments can be applied outside of copyright law to include harmful content generated with SD.
Well they don’t own the LAION dataset, which is what their image generators are trained on. And to sue either LAION or the companies that use their datasets you’d probably have to clear a very high bar of proving that they have CSAM images downloaded, know that they are there and have not removed them. It’s similar to how social media companies can’t be held liable for users posting CSAM to their website if they can show that they’re actually trying to remove these images. Some things will slip through the cracks, but if you show that you’re actually trying to deal with the problem you won’t get sued.
LAION actually doesn’t even provide the images themselves, only linking to images on the internet, and they do a lot of screening to remove potentially illegal content. As they mention in this article there was a report showing that 3,226 suspected CSAM images were linked in the dataset, of which 1,008 were confirmed by the Canadian Centre for Child Protection to be known instances of CSAM, and others were potential matching images based on further analyses by the authors of the report. As they point out there are valid arguments to be made that this 3.2K number can either be an overestimation or an underestimation of the true number of CSAM images in the dataset.
The question then is if any image generators were trained on these CSAM images before they were taken down from the internet, or if there is unidentified CSAM in the datasets that these models are being trained on. The truth is that we’ll likely never know for sure unless the aforementioned trials reveal some email where someone at Stability AI admitted that they didn’t filter potentially unsafe images, knew about CSAM in the data and refused to remove it, though for obvious reasons that’s unlikely to happen. Still, since the LAION dataset has billions of images, even if they are as thorough as possible in filtering CSAM chances are that at least something slipped through the cracks, so I wouldn’t bet my money on them actually being able to infallibly remove 100% of CSAM. Whether some of these AI models were trained on these images then depends on how they filtered potentially harmful content, or if they filtered adult content in general.