California AG to OpenAI: Harm to Children Will Not Be Tolerated
oag.ca.gov
oag.ca.gov
Completely possible, compatible, and probably mostly effective. The only question is whether there's the political will and the societal tolerance.
Work out how much state and corporate power we would need to come to bear to enforce the vision of the world you are advocating.
Or, bravo on the HN troll. This is perfectly triggering for this crowd.
Hell, you can download scans of old porn magazines on archive.org.
I don't think that would be effective, and I definitely don't think there is a political will.
2/5s of Gen Z can't even navigate between folders on a computer, according to studies, with CS instructors saying they need to introduce directories as a concept. Yes, it will.
2/5ths of Gen Z not currently knowing how to navigate between folders in a hierarchical file system doesn't mean that 2/5ths of Gen Z is incapable of learning how to navigate between folders in a hierarchical file system.
The gap between present knowledge and capacity doesn't exist due to incompetence, it exists due to lack of motivation, and rest assured, once teenage boys discover how great boobs are, they'll have all the motivation they need to learn how to use technology to achieve the goal of viewing and appreciating that kind of material, pretty much regardless of how many restrictions you put in their way.
They'll make their own in a paint app or with ASCII art in a notepad, push come to shove, but that won't be necessary. It's hilarious to think that states have the slighest hope of winning the whack-a-mole game of trying to sue every single website on the entire internet that hosts any pornographic content.
How many people do you know whose first introduction to porn was "that's sounds cool, let me look that up?" Versus discovering it by accident, exactly what a law would effectively prevent?
I'm not opposed to measures that try to make it harder for kids to get at porn. But it's going to be pretty much impossible to do that as long as we want adults to have unimpeded access to it, and we do.
One option would be to put the responsibility on the distributor, and not give them the option of a "you must be 18 to enter" fig leaf: if a kid is found in possession of porn that can be traced to your web site, you face serious fines and possible imprisonment. But that would basically eliminate porn sites for everyone, since it would be impossible to prevent it from finding its way to kids, and we can't have that.
But again, none of that will be necessary. Rule34 websites, imageboards, video game rendering glitches... the list goes on and on.
It's truly delusional to believe that it's possible to exhaustively censor imagery of the human body from the most motivated seeking/viewing demographic on the planet.
Are you by chance a woman? Were you homeschooled? Do you not have memories of being a teenage guy in the US public school system? I don't mean to denigrate you, I just cannot fathom how anyone can possibly credibly believe that it's even slightly realistic to keep teenage boys completely unaware of what boobs look like in the real world.
Oh, man. You just reminded me of teenage me trying to make something out in scrambled porn on cable TV after midnight. You could kind of see the thrusting, but it was barely stimulating. Those were tough times, let me tell you.
I was literally looking up Zero Suit Samus fan art on my Wii's web browser when parents put parental control software on the computer.
When I was grounded from all computer usage, I'd pull up Catie Minx photoshoots on a first-generation Kindle, with the e-ink display that had image update latency measured in seconds.
All of my friends would go to these extents and more to access this kind of material. I was introduced to dumpster diving by a friend for this reason. I had another friend who got caught for shoplifting adult magazines from a bookstore.
I don't think the motivation and ingenuity in seeking adult content is generationally restricted or particularly unique at all, it seems to be true for every guy I know, including those who were born in the 50s, 60s, 70's, 80s, 90s.
I genuinely cannot comprehend the thought process of people who truly believe that a few targeted lawsuits at major porn sites is going to make a dent in what might be one of the strongest behavioral drives hardcoded into the human genome.
I'll also add that regardless of any access to web or printed materials, boys will get interested in boobs and seeing more of them by seeing them in the real world, on real women.
There is like no stronger instinct in human DNA than for males to be attracted to females. If it weren't that way our species probably wouldn't have survived down the millenia.
Are they capable of using computers to achieve their goals? Yes. Does that require navigating folders? No.
EDIT0: I remember writing a 4 line program to crash the school computer (recursively executing itself) on notepad, and I'm pretty sure I didn't know what folders and directories were.
I can also attest to many of my coworkers not understanding what file extensions are, while being over 40 and working in aeronautics.
In other words, exactly what their plan is.
Why not just add whitelisted parental controls on childrens phones and be done?
Allowlist-based parental controls already exist. Most parents never set them up because it's a pointless battle, their teenagers will just another device that isn't controlled by the parent.
Where there is a will there is a way.
Now, notice I did NOT say that there actually is sufficient will, at the moment.
Papers Please.
Here is the messed up thing, there is a good chance that one could make a pretty accurate guess on your age just based on behavior.
The digital age needs its own "steal this book"
legally unconstitutional to enforce
Same for many other movies from that era.
Oh wait. That was Hanna Barbera, Disney, Looney Tunes, and the like.
Later childhood after watching the re-re-repeats of 1960's shows, was learning about hatred, racism, and what the 'normies' would do to you, no matter what you did or didn't do. And, Magneto was right.
Children are exposed to liars, to fools, to the stupid, to the unwise, to the corrupted in malice, to the weak in will... Since a very young age.
«Exposing» such frequent faults to children does not «normalize» it. It may, if the child is in a generally bad environment - but then, the target is the core environment, not the outer experience.
Not really easy, I'd say.
> trade-offs
If you mean barring general access, than your evaluation is debatable.
But especially:
if the concern is with some forms of possibly "certified traumatic" depictions (and I believe this would raise a hell of a needed research and nuances revelation),
why not just attacking the "certified traumatic" depictions?
That's my gripe. Vixen is veeeery different than some other websites under the guise of "porn" - they're going to ruin it for all of us
There's a known, repeatable failure mode in these engines that anyone who's worked with them for more than a couple hours can tell you: when a conversation goes on too long, it "feeds back" on itself. I don't fully understand the mechanism (I believe it's partly to do with the model's attention mechanism getting saturated and newer content dominating over older content, and it's related to the "jailbreak" solution where you hit the machine with so much text that it "forgets" the directives it's been given by its creators when it booted up). But the end result is that over time, the conversation centers on the more recent topics, ideas, and tones over the initial configuration. That's one of the reasons you can "trick" these models into being racist by talking like a racist, and so on.
Reading the excerpts I've read from the chat history Raine had with his session, it seems pretty clear that it's gone on so long that the session is "reflecting" his own writing and mood back at him. And that's the heartbreaking part: it's, in essence, coaching him to end his life because he's been talking about ending his life for so long in the conversation that it's saturated the model.
It's this poor young man's own pain reflected back on him.
It's like an inhuman instinct, ingrained into the model by the pre-training process. To the model, the context is the world it sees - and it will always try to match its outputs so that they fit in with what it sees. In many ways, this context-matching process is at the very foundation of an LLM's behavior.
And the more context there is, the more constrained the space of "consistent continuations" becomes, and the harder it may become to break a model out of a rut it got stuck in.
This is what gives power to few-shot prompts. The model is primed to look at prior demonstrations, and act in a way that's consistent with what was demonstrated to it. Self-consistency works in your favor there. But it's also what powers a lot of unwanted LLM behaviors. Like poor multi-turn instruction following - when the model's self-consistency begins to dominate the conversation, and its own prior actions begin to have more weight in its future behavior than the instructions added by the user (hi Gemini).
This also means that you can totally "boil the frog" by shifting the vibe of the context over time. A well trained "harmless, honest and helpful" chatbot would never tell the user "you should try fentanyl"! But keep the conversation going for long enough, hit on the right themes, and that innate consistency drive may begin to overpower the "HHH" training. At a certain point, if the vibe of the conversation permits, the LLM might actually say "you should try fentanyl" to its user, training be damned.
Another tricky thing is that you can totally make a frog that boils itself! For example, if your model (hi GPT-4o) has a sycophancy bias? Then with every time it agrees with the user in an over-the-top sycophantic fashion, the strength of "the AI is a total sycophant" in the context grows. And the AI wants to match the context! The innate sycophancy bias and the sycophancy bias added by the self-consistency drive then feed into each other. Until the combined might of the two overpowers the "HHH" training, and the AI tells you that your idea to add some paint thinner into your herbal tea is a brilliant insight!
while giving old tech a pass
The "think of the children" crowd should not be given a single inch. Nothing good ever came from it, and by now, I believe that nothing ever will.
Well, reason on it, and you'll see that reason will confirm the instinct.