(AI images) (Legality) "Man Arrested for Producing, Distributing, and Possessing AI-Generated Images of Minors Engaged in Sexually Explicit Conduct"

I wasn’t sure of which forum to post this in, but this one seems fine, because it’s about internet, technology, and law.

I am concerned. About the indictment, and about the language used in their press release:

“As alleged, Steven Anderegg used AI to produce thousands of illicit images of prepubescent minors, and even sent sexually explicit AI-generated images to a minor,” said Principal Deputy Assistant Attorney General Nicole M. Argentieri, head of the Justice Department’s Criminal Division. “Today’s announcement sends a clear message: using AI to produce sexually explicit depictions of children is illegal, and the Justice Department will not hesitate to hold accountable those who possess, produce, or distribute AI-generated child sexual abuse material.”

Indictment: https://www.justice.gov/opa/media/1352606/dl?inline

“Government’s brief in support of detention”: https://www.justice.gov/opa/media/1352611/dl?inline

Press release: Office of Public Affairs | Man Arrested for Producing, Distributing, and Possessing AI-Generated Images of Minors Engaged in Sexually Explicit Conduct | United States Department of Justice

Where I stand on this is simple.
Sending those images to a minor (I think he was 14 years old) is wrong.
Generating those images is ethically OK and free speech, and it helps that they do not allege that the outputs resemble a real, identifiable minor person.

It seems that he wasn’t another creep who was using photos of children that he knows or sees around him- I can understand why a person like that should pay something for what he did (but not incarceration, that’s crazy).
If I had children, and some guy was taking photos of my kids, I would want his balls destroyed by a rail gun.

But, it looks like all this guy did is generate a ton of images of young-looking human characters doing very explicit things, and then sent some of them to a teenager on Instagram, a teenager who did state his correct age to the adult (he didn’t lie about his age like many teens probably do)

Instagram’s image classification is what flagged his messages for review. Obviously, Meta/Facebook/Instagram have their own models to help identify potential underage nudity, leaning toward the younger ages (it’s 1000x easier to correctly detect a toddler than a 17 year old girl). It is still hilarious, and insidious, that a human reviewer saw what are obviously fake AI images, and decided to escalate it so it would be reported to law enforcement as if it were CP.

I like “oneshota” (please correct me if I’m using the wrong word), and I have prompted Stable Diffusion for mild scenes, and done 3D renders of fictional women with fictional younger ones, and I occasionally favorite both 2D and 3D renders of this kind on platforms like Pixiv, so I am increasingly concerned.
I still feel “safe”, but I am disgusted by the fact that I cannot accept donations, on a platform like Pixiv Fanbox or SubscribeStar, without the risk of being arrested because my “softcore” images are too realistic or something. I also feel that making donations to such artists is potentially high risk, for the same reason.

It doesn’t help that I’ve caught news media AND POLICE misidentifying images as “AI” when they are actually 3D renders. I checked some cases in UK and in the US that were about “AI” and ended up being about “oh he did some Stable Diffusion too but we’re adding these 3D renders to the charges”, in one case I was able to find an original image that landed a man in jail (one of many he was charged for), because the filename in the legal document was enough to find the original image on rule34 dot xxx because the relevant tags were in the filename. I found it, I saw it, and to me, it was a nothing-burger. This man’s charges included AI renders, and similarly nothing-burger 3D renders like that one, with a woman and a boy who are obviously not real people.

TOO LONG DIDN’T READ:

It looks to me like computer-generated images are on the chopping block, in the United States!

Please let me know if it’s true or not, so that I can be anxious about one less thing, or start reviewing my security posture.

3 Likes

He was charged for sharing them with an actual minor, which is a crime no matter what they depict.

The actual legality of such images, like with adult pornography, is ultimately still in a ‘gray area’, whereby such images are presumed legal unless ‘proven’ to be ‘obscene’ in a court of law.

It’s also very likely that he was using photos of real children and the DoJ releases just didn’t showcase that fact. There was another case involving a soldier using identical means to produce images but specifically included in the prompts terms that referenced photographs of real children.

5 Likes

I thought he also received child porn related charges, but if he really didn’t, that’s a relief. I actually caught wind of this from a Guardian article that flat-out said that the FBI charged him because of the “AI child porn” (I’m paraphrasing, but, yes)

Frankly I get upset about this repeated mishandling of “AI child porn” because, I feel that the moral orel borg (or, the borg of virtue signalling) will soon be encroaching on the domain that I actually enjoy, which is 3DCG.

I am disgusted that there is a chance my Blender project files are “illegal”

And in some countries, that is already true. People losing their freedom because of computer graphics. And it just hurts me, seeing artists in general, making beautiful art (in any medium, especially traditional ones), being arrested in some European countries because they accidentally made “”“”“”“”“”“”“”“”““child porn””“”“”“”“”“”“”“” when they declared that the cute anthropomorphic blue jay they drew with their Wacom tablet is a 16 year old high school student. Oh and, that’s an approximate example! This was a French female artist I followed. Think of the taxes wasted on that whole process. The panic she endured for over a month as she went through legal process with her government-appointed lawyer and was nearly jailed over her own art- no, nearly jailed over her own words describing her own art. Imagine the complete loser who decided to report her, starting with sending her Inkbunny/FurAffinity account URLs to her country’s authorities (which lead to legal process that divulged IP address, which is useful enough, most of the time)

Technical measures like this are ending the freedom of artists, and their audience, alike.

Technical measures can also defend people.
Tor to raise internet privacy, Monero to raise payment privacy.

Tor brings less friction, IMO… but I totally understand why people cannot simply “get work done” with Monero, because trading it for fiat is intentionally made difficult by exchanges (they simply refuse Monero, because regulators hate systems that bring privacy where there is no privacy).

1 Like

I’m curious about one more thing- I know no one here can read the minds of such defective people (well maybe the old timers can!),
but,
why are headlines so loaded, and so loaded in favor of painting AI child porn as the ultimate reason for arrests, charges, etc.?

This is not an isolated incident, Ive seen it from US local news, UK news, Aussie news, etc. etc. etc.

But even the DOJ, the FUCKING DOJ, has their loaded headlines and their unelected officials saying it’s because of the AI images, WTF

See, I quoted it in the OP-

“As alleged, Steven Anderegg used AI to produce thousands of illicit images of prepubescent minors, and even sent sexually explicit AI-generated images to a minor,” said Principal Deputy Assistant Attorney General Nicole M. Argentieri, head of the Justice Department’s Criminal Division. “Today’s announcement sends a clear message: using AI to produce sexually explicit depictions of children is illegal, and the Justice Department will not hesitate to hold accountable those who possess, produce, or distribute AI-generated child sexual abuse material.”

And yes, I know it MIGHT be because he used actual photos as inputs, but the government still has to prove it, AND THEY FLAT OUT SAID THAT HE HAD LIKE 12,000 IMAGES OF AI “”“CP”“”, I don’t see how they could PROVE ALL OF THAT SH*T, right? (automatically going through his ComfyUI/EasyDiffusion/Oobabooga prompt history would not count as useful in legal process, right? Because you have to prove that the outputs themselves resemble a real child)

I really hate these people

2 Likes

If images generated are not evidence that what is depicted actually occured, they cannot be charged as CSAM in the US.

If the images show no hardcore conduct, they cannot be found obscene under Miller. Nudity alone doesn’t suffice.

One needs to be careful about what they send to a minor.

Consequently, there is no test for content that can be found inappropriate for a minor.

Ginsburg v. New York (1968)

3 Likes

Looks like another case of a dude who was being a danger to actual minors and the government used what they had available as a means to get him the most years. I am not surprised at all.

1 Like