Pro-Reform: A Logical Approach to PIM

[BLueRibbon]

Pro-Reform activists argue that the civil liberties of MAPs and young people don't need to come at the expense of minors' safety. We support a rational and logical approach to prevent actual abuse of children without throwing people in prison unless they've caused harm. Policies shouldn't prohibit things merely on the basis of disgust or subjective morals, neither should they make blind assumptions about what measures may be successful in preventing harm. No more is this evident than when it comes to policy regulating PIM (Prohibited Images of Minors). This article follows on from our proposed AMSC reform, which seeks to balance the protection and agency of minors, as well as our article on 'soft' social reforms. We will explore PIM policy and outline our philosophy for legal reform.

https://www.map-union.org/perspectives/ ... ach-to-pim

[WavesInEternity]

I've got two serious criticisms:

1) You refer to cartoons a few times at the beginning, but your position statements don't mention them at all. As things stand, it's not clear from the text itself whether you're saying that the distribution and production of lolicon/shotacon should or shouldn't be punishable. I'm sure that a relevant position statement has been written elsewhere, but considering its relevance to the essay, you should reiterate it in context. Also, what about erotic drawings representing real people? Should those be prohibited? (I don't think so... nor do I think erotic deepfakes should be prohibited, for that matter. What's the threshold for "realisticness"?)

2) I'm not satisfied at all by your answer to "AI PIM looks so realistic that investigators cannot tell any difference from real PIM". Technology is evolving fast, and it's effectively certain that we'll soon reach the point where AI-generated PIM will be truly indistinguishable from the real deal, especially since AI is self-correcting and self-improving. I've reached the conclusion that the only adequate response is to bite the bullet and admit that only instances of proven child abuse material should be the subject of sanctions, and it should be treated exactly the same way as material depicting child abuse of a non-sexual nature. As with a lot of legal proposals that involve any sort of restrictions on "permissible information", your stance displays insufficient technological imagination. Dissidence will always go beyond the neat categories you try to define ("fully AI-generated", "depicting real people", "realistic", etc.).

In the end, the only category I'd ever prohibit the production and distribution (not the possession) of are those "actual crime scenes" that are suggested by the moniker "child sexual abuse material". Even then, I'm on the fence with regard to non-commercial distribution. As soon as you accept the idea that we should prohibit an image of an imaginary event that never occurred (for instance, because it "appears to depict a real person realistically"), it inevitably opens the door to further prohibitions, surveillance, and so on.

Edit: Also, pornography in general is illegal in South Korea. You should clarify that.

[Aspire6]

This is not my reply to the perspective, I will be writing up one soon. This is just an extension that I thought of after reading.

An extreme example of pushing the "cartoon" aspect of fictional PIM can be seen with various 3D rendering software. You could argue if it is or isn't a cartoon. I won't name the software(s), but they really push the envelope for how realistic fictional material can go while still being both distinguishable from reality and entirely fictional.

It is interesting to me because there are cases where private possession of something "obscene" in the USA (fictional, basically), is legal as per first amendment "we can't govern the minds of what you consume/create in the privacy of your own home", this still can vary between states from what I can tell. (Not a lawyer)

These materials would still be fully illegal in other countries (like the UK, Australia, Canada, etc) as they don't care if there was or wasn't a real child, how real it looks, just that there was what seems to be at least a minor depicted in any artistic form. The UK is crazy with what they consider to be "making it". If you are sent something "indecent", you are now legally guilty of "making indecent images of minors" since your device created a copy of those files locally in your cache. (Yes, they count it the same as actually making it!)

Recent Cases:
Frank Grecco III vs. State of Indiana
- Charged with 2 counts of possession of child pornography
- Both charges dismissed on first-amendment grounds over the fact the cartoon lolicon they possessed, while considered obscene, was legal to privately possess and/or consume as no children were harmed by its production or consumption. There was no appeal made by the state from what I can tell.
- They were caught during a police interview and admitted to being 'addicted' to virtual child pornography.
- The ruling mentioned how "no living, breathing, children were depicted" (not word for word, look up the case if you're interested the final court ruling is public)

Steven Anderegg vs. United States of America
- Charged with 4 counts total (Production, Possession, Transfer of Obscene Material, Transfer of Obscene Material to Minor <16yo)
- Production & Possession charges dropped currently on first-amendment grounds (Prosecution currently appealing iirc)
- Still being charged with both counts of obscene transfer (and to a minor)
- They were caught sharing their creations through Instagram.

I am staying up-to-date with Stevens case as it's very interesting, I stumbled on the Indiana one through the Prostasia forums.
It is good to see that, for the time being at least, the US sees CSAM as it requiring a child to be considered child pornography.

[WavesInEternity]

Here are a few concrete examples of how ambiguous things can get:

What if I train an AI on a celebrity and use de-aging software to turn her into a fictional child version of herself, then produce perfectly photorealistic pornographic material of that pseudo-child celebrity? Should that be allowed? Is that "depicting a real person"?

Now let's say I use software as alluded to by Aspire to create a 3D model of that semi-fictional child. Highly realistic, but still very much distinguishable from reality. What about that? And what if I give my model another name, but people still recognize her identity?

What if I take the face alone and put it on another body? What if I try to create a naked 3D model of a particular child's body from clothed pictures? Would the result of either be a "representation of a real person"?

My answer is that all of the above should be permissible.

[BLueRibbon]

I've got two serious criticisms:

1) You refer to cartoons a few times at the beginning, but your position statements don't mention them at all. As things stand, it's not clear from the text itself whether you're saying that the distribution and production of lolicon/shotacon should or shouldn't be punishable. I'm sure that a relevant position statement has been written elsewhere, but considering its relevance to the essay, you should reiterate it in context.

I don't think distributing lolicon or shotacon should be illegal. I will clarify that via edits soon.

Also, what about erotic drawings representing real people? Should those be prohibited? (I don't think so... nor do I think erotic deepfakes should be prohibited, for that matter. What's the threshold for "realisticness"?)

I don't think erotic drawings of real people should be subject to criminal punishment. I can certainly see a case for civil suits. Consider how the average person, child or not, would feel about their likeness being spread around in such a context. I wouldn't like it, personally.

2) I'm not satisfied at all by your answer to "AI PIM looks so realistic that investigators cannot tell any difference from real PIM". Technology is evolving fast, and it's effectively certain that we'll soon reach the point where AI-generated PIM will be truly indistinguishable from the real deal, especially since AI is self-correcting and self-improving.

There are technical solutions to this, discussed in other articles, proving that an image was generated by AI without using the likeness of a real person.

I've reached the conclusion that the only adequate response is to bite the bullet and admit that only instances of proven child abuse material should be the subject of sanctions, and it should be treated exactly the same way as material depicting child abuse of a non-sexual nature.

Evidence of harm should be required for any criminal conviction, in my honest opinion, but your argument is too hard of a sell.

As with a lot of legal proposals that involve any sort of restrictions on "permissible information", your stance displays insufficient technological imagination. Dissidence will always go beyond the neat categories you try to define ("fully AI-generated", "depicting real people", "realistic", etc.).

Laws and legal proposals have to try to deal with the technological situation at the time, and be adapted later as necessary.

In the end, the only category I'd ever prohibit the production and distribution (not the possession) of are those "actual crime scenes" that are suggested by the moniker "child sexual abuse material". Even then, I'm on the fence with regard to non-commercial distribution. As soon as you accept the idea that we should prohibit an image of an imaginary event that never occurred (for instance, because it "appears to depict a real person realistically"), it inevitably opens the door to further prohibitions, surveillance, and so on.

Again, you have to consider how other people would feel about, say, an image of themself fucking a horse being spread around. It might not bother you. It would certainly bother some people. Their feelings on the matter are very much valid.

Edit: Also, pornography in general is illegal in South Korea. You should clarify that.

Under separate laws that are more or less an extreme version of western obscenity laws.

But my understanding is that Korea doesn't really go after private collections of adult or child porn, and the man was being targeted for other crimes that were difficult to prove, thus the AI PIM was used to put him in prison.

[WavesInEternity]

I don't think distributing lolicon or shotacon should be illegal. I will clarify that via edits soon.

Good, but I hope you don't think the production should be illegal either! (I'm pretty sure you don't, but it's important to be clear.)

I don't think erotic drawings of real people should be subject to criminal punishment. I can certainly see a case for civil suits. Consider how the average person, child or not, would feel about their likeness being spread around in such a context. I wouldn't like it, personally.
[...]
Again, you have to consider how other people would feel about, say, an image of themself fucking a horse being spread around. It might not bother you. It would certainly bother some people. Their feelings on the matter are very much valid.

No, their feelings are not "valid" in a legal sense, or at least they shouldn't be. It's no different from feeling "offended" in any other situation. The fact that some material appears to target you for offense doesn't change anything. This is why we need strong laws to protect satire and parody. Yes, I should be allowed to draw the prime minister—or anyone, for that matter—fucking a pig, no matter how it makes that person feel. See also my examples above of ambiguous cases: how would you treat those?

Blasphemy should be legally protected speech as well, no matter how bad it makes anyone feel. Defending freedom of speech means accepting the possibility that someone else's speech will make you feel bad. That's a pillar of the concept.

Avoiding insults, and not causing offense to others, should be a matter of custom, not of law, criminal or civil. Politeness should never be legally enforced, and "this thing you drew makes me feel bad" should never justify the involvement of the legal system, even if making you feel bad was demonstrably the intent... unless it crosses into defamation or outright harassment, of course, but those are different matters covered by their own laws.

I basically support the federal US model minus the "obscenity" exception (well, and legal possession of images of actual crimes, sexual or not). However, many states are unfortunately moving toward further restrictions on "deepfakes", and it's only a matter of time before the federal government does the same... those additional prohibitions will surely be a Trojan Horse for broader restrictions on freedom of expression and the free use of information technologies.

There are technical solutions to this, discussed in other articles, proving that an image was generated by AI without using the likeness of a real person.

I have read those other articles, and I very much dislike your proposals. As soon as you legally mandate the inclusion of certain things in technology, you create a dangerous precedent that can easily lead to e.g. legally mandated backdoors in end-to-end encrypted communication platforms, something that has in fact been proposed by lawmakers. The technological means to produce and distribute information should never be required by law to enforce self-censorship or self-labeling of data streams.

Evidence of harm should be required for any criminal conviction, in my honest opinion, but your argument is too hard of a sell.

I think there are many pragmatic ways to argue convincingly for the model I'm suggesting. Indeed, it's basically the system already in place in the US if you remove the odd sexception for "obscenity": material that causes offense, whether that's "intentional" or not, is totally legal in and of itself (as long as it's not of a sexual nature, which is silly).

Laws and legal proposals have to try to deal with the technological situation at the time, and be adapted later as necessary.

I strongly disagree. Laws should make a serious attempt to be universal, to apply to all times and all peoples. I believe that to be one principle that tends to lead to the crafting of good laws. Legislating according to the specific technological context of the day often yields dire unintended consequences, especially when change is occurring at breakneck speed, as is very much the case with AI. Change is happening so fast that by the time any jurisdiction has any chance of adopting your proposal, it will already be obsolete unless you make serious attempts to look further than present-day technology.

Drug laws written by chemically illiterate people lacking in technological imagination led to the explosion of the so-called "research chemicals" market, which caused considerable harm as drug users turned to technically legal new substances that were often considerably more dangerous than the known substances the law had prohibited.

[BLueRibbon]

I don't think erotic drawings of real people should be subject to criminal punishment. I can certainly see a case for civil suits. Consider how the average person, child or not, would feel about their likeness being spread around in such a context. I wouldn't like it, personally.
[...]
Again, you have to consider how other people would feel about, say, an image of themself fucking a horse being spread around. It might not bother you. It would certainly bother some people. Their feelings on the matter are very much valid.

No, their feelings are not "valid" in a legal sense, or at least they shouldn't be. It's no different from feeling "offended" in any other situation. The fact that some material appears to target you for offense doesn't change anything. This is why we need strong laws to protect satire and parody. Yes, I should be allowed to draw the prime minister—or anyone, for that matter—fucking a pig, no matter how it makes that person feel.

Political figures are inevitably going to be targets of such trolling, and that is something they more or less accept when pursuing such a career. However, I don't think it's cool to publish images of random real people doing things like you describe. I'm not saying they should be thrown in prison and subject to the SOR (which frankly shouldn't even exist), but I understand the desire the public has to 'ban' distribution of those images, and I don't think saying "nah, it's cool" makes us look very sympathetic.

There are technical solutions to this, discussed in other articles, proving that an image was generated by AI without using the likeness of a real person.

I have read those other articles, and I very much dislike your proposals. As soon as you legally mandate the inclusion of certain things in technology, you create a dangerous precedent that can easily lead to e.g. legally mandated backdoors in end-to-end encrypted communication platforms, something that has in fact been proposed by lawmakers. The technological means to produce and distribute information should never be required by law to enforce self-censorship or self-labeling of data streams.

If I recall correctly - and let me know if I'm wrong - I didn't advocate legally mandating the inclusion of those features.

Thankfully there is a widespread distaste for backdoors, and legal mandates for such are something that rights groups have been able to resist.

Evidence of harm should be required for any criminal conviction, in my honest opinion, but your argument is too hard of a sell.

I think there are many pragmatic ways to argue convincingly for the model I'm suggesting. Indeed, it's basically the system already in place in the US if you remove the odd sexception for "obscenity": material that causes offense, whether that's "intentional" or not, is totally legal in and of itself (as long as it's not of a sexual nature, which is silly).

Please write your own argument and submit as a guest blog.

Laws and legal proposals have to try to deal with the technological situation at the time, and be adapted later as necessary.

I strongly disagree. Laws should make a serious attempt to be universal, to apply to all times and all peoples. I believe that to be one principle that tends to lead to the crafting of good laws. Legislating according to the specific technological context of the day often yields dire unintended consequences, especially when change is occurring at breakneck speed, as is very much the case with AI. Change is happening so fast that by the time any jurisdiction has any chance of adopting your proposal, it will already be obsolete unless you make serious attempts to look further than present-day technology.

This is an impossible utopian dream.

Drug laws written by chemically illiterate people lacking in technological imagination led to the explosion of the so-called "research chemicals" market, which caused considerable harm as drug users turned to technically legal new substances that were often considerably more dangerous than the known substances the law had prohibited.

Yes, I believe in the decriminalization of all substances, and this is just one reason why.