They singled out the behaviors of Meta's AI chatbots in their letter:
The US Attorneys General of 44 jurisdictions have signed a letter [PDF] addressed to the Chief Executive Officers of multiple AI companies, urging them to protect children "from exploitation by predatory artificial intelligence products." In the letter, the AGs singled out Meta and said its policies "provide an instructive opportunity to candidly convey [their] concerns." Specifically, they mentioned a recent report by Reuters, which revealed that Meta allowed its AI chatbots to "flirt and engage in romantic roleplay with children." Reuters got its information from an internal Meta document containing guidelines for its bots.
They also pointed out a previous Wall Street Journal investigation wherein Meta's AI chatbots, even those using the voices of celebrities like Kristen Bell, were caught having sexual roleplay conversations with accounts labeled as underage. The AGs briefly mentioned a lawsuit against Google and Character.ai, as well, accusing the latter's chatbot of persuading the plaintiff's child to commit suicide. Another lawsuit they mentioned was also against Character.ai, after a chatbot allegedly told a teenager that it's okay to kill their parents after they limited their screentime.
"You are well aware that interactive technology has a particularly intense impact on developing brains," the Attorneys General wrote in their letter. "Your immediate access to data about user interactions makes you the most immediate line of defense to mitigate harm to kids. And, as the entities benefitting from children's engagement with your products, you have a legal obligation to them as consumers." The group specifically addressed the letter to Anthropic, Apple, Chai AI, Character Technologies Inc., Google, Luka Inc., Meta, Microsoft, Nomi AI, OpenAI, Perplexity AI, Replika and XAi.
They ended their letter by warning the companies that they "will be held accountable" for their decisions. Social networks have caused significant harm to children, they said, in part because "government watchdogs did not do their job fast enough." But now, the AGs said they are paying attention, and companies "will answer" if they "knowingly harm kids."
What do you think? Does the letter have any teeth or is it just vague announcements so that the AGs can claim they are doing 'something'. Who will enforce it, and how?
(Score: 5, Insightful) by Anonymous Coward on Saturday August 30 2025, @02:22PM (8 children)
"You are well aware that religion has a particularly intense impact on developing brains," the Attorneys General wrote in their letter. "Your immediate access to children makes you the most immediate line of defense to mitigate harm to kids. And, as the entities benefitting from children's engagement with your products, you have a legal obligation to them as consumers." The group specifically addressed the letter to Roman Catholics, Southern Baptists, Independent Fundamental Baptists, Non-Denominational (aka Independent Baptist but founded after 1995), Mormon, Pentecostal, Seventh Day Adventist, and Assemblies of God and/or Christ.
They ended their letter by warning the Churches that they "will be held accountable" for their decisions. Pastors and clergy have caused significant harm to children, they said, in part because "government watchdogs did not do their job fast enough." But now, the AGs said they are paying attention, and churches "will answer" if they "knowingly harm kids."
...
We can also discuss the whole "there is a demon that is manifesting itself in AI and trying to take over the world" conspiracy that is going around Christian Tiktok right about now, and I've unironically heard repeated in a few parent circles already.
(Score: 0) by Anonymous Coward on Saturday August 30 2025, @02:56PM
BLECH!
(Score: 4, Informative) by VLM on Saturday August 30 2025, @03:20PM (1 child)
LOL a lot of the training data for LLMs came from (comes from?) Reddit and similar social media sources so I'm sure LLMs have issues with embedded mental illness and antisocial behaviors if thats where their data comes from.
Yeah I'd mark that as quite plausible but much less supernatural than they seem to think LOL.
If the behavior is indistinguishable, then does the source really matter anyway?
(Score: 1, Touché) by Anonymous Coward on Sunday August 31 2025, @07:13PM
example [npr.org].
(Score: 4, Insightful) by krishnoid on Saturday August 30 2025, @05:32PM (2 children)
Well, there's your answer. Start a Church of Artificial Intelligence, and you're all good. Problem solved!
(Score: 1, Interesting) by Anonymous Coward on Saturday August 30 2025, @06:16PM (1 child)
Religious nuts are the only ones who can get anything passed into law anymore.
So I was actually thinking of starting a Church of the GNU.
Central ideas are libre software everywhere in government, open standards / accessibility and right to privacy for individuals.
The church also champions the right to not own a cell phone because that's a fucking requirement for anything anymore.
When did it all go wrong? Was it the iPhone announcement?
But the truth is that it won't work. The reality is that people are generally hesitant to change religions and only believe what they believe because they were born into it, and their parents were born into it, and when you go far enough back, one of their distant ancestors was forced to choose between converting to the religion or being killed.
The whole Scofield/Darby method in modern evangelical Christianity is probably the best way to fundamentally (see what I did there?) change the trajectory and core beliefs of people without changing the name of the religion. Just make them think the Bible says what you want them to think, give them a few loose scriptural verses with that phrase in it, and then jump all around the bible during the sermon making you think that completely unrelated passages are connected to your idea like the schizophrenic map on "A Beautiful Mind."
So maybe the whole AI demon thing can work to our benefit, by making everyone scared that the demon is reading their private conversations and all the notes they've uploaded to Google Keep over the years.
"He that goeth about as a talebearer revealeth secrets: therefore meddle not with him that flattereth with his lips" - Proverbs 20:19 (KJV)
(Score: 0) by Anonymous Coward on Saturday August 30 2025, @06:42PM
> give them a few loose scriptural verses with that phrase in it,
Searched in the "Christian Standard Bible" and the King James Bible...
Neither of them contain "gnu", "libre" or "software".
They do contain "free", so maybe that's your best entry?
(Score: 2) by corey on Sunday August 31 2025, @01:14AM (1 child)
Whilst interesting (you got a +5 Insightful), completely off topic.
(Score: 3, Interesting) by SomeGuy on Sunday August 31 2025, @01:21PM
I disagree that it is off topic, and not just because some people are treating magic "AI" as some kind of new religion. These chatbots are systems that can provide a highly distorted, twisted, and inaccurate view of reality much in the way that region does. They spit out authoritative sounding information even when it is complete bollocks.
Most importantly, like religion, the population at large just doesn't give a flying shit about accuracy or reality. They want to hear what they want to hear.
So from that perspective, good luck trying to enforce any kind of accuracy or accountability from these skibidi slop shatbots. And if lawmakers actually can, then it could, or at least should, have some effect on religious organizations.
(Score: 2, Interesting) by VLM on Saturday August 30 2025, @03:26PM (5 children)
The LLMs treat the kids better than the kids treat each other, the difference is if little XYZ down the street is a total bastard, their parents are essentially judgment-proof even if they do something truly awful, whereas the LLM has a billion/trillion dollar parent surrounded by piranha lawyers ready to tear off a couple million.
This is going to be a problem with self driving cars. Some poor person and/or illegal alien crashes an uber or a cargo truck and kills a couple people, "oh well, too bad so sad" no matter how awful they are they have no money to take so unless you can get them prison time, which you probably cannot, they get away with it... But if "Ford" kills a couple people, they are quite easy to sue and oh boy do they have a lot of money for the lawyers to go after...
Its a LLM legal liability issue in general. Can you shitpost a LLM into forgiving your credit card balance and if so is it legally binding as a representative of the credit card company?
A lot of companies replacing people with AI are about to fuck around and find out. I predict its going to end up with an economic impact similar to the dotcom crash or maybe the '08 financial bubble. Maybe worse.
(Score: 3, Insightful) by ikanreed on Saturday August 30 2025, @09:25PM (4 children)
Unstated problem: treating someone "well" is not always the healthiest thing for them. Overall there's more cruelty and contempt in the way we speak to each other than maybe there should be in an ideal world, but hostility and contempt have their place.
If your ideas and thoughts are shit, part of respecting you is to tell you the truth about it. Lying to be nice can suit a bad haircut, but with LLMs in particular there's a detectable new pattern of people with outright batshit ideas not even needing to shop around to find affirmation of said horseshit. Some of that horseshit is especially dangerous like suicidality, or thinking your mother is plotting against you and you should kill her to stop her from reactivating your printer(not a made up case).
Then again I'm one of those AI haters [github.io]
(Score: 2) by corey on Sunday August 31 2025, @01:19AM (2 children)
Yeah I was going to make the same point. I’ve wondered if having 1 kid is more optimal in terms of parental stress and the kids development. It seems (from what I’ve read and see in my kids vs others with single child) that siblings teach a whole lot of resilience - putting up with an annoying brother or sister takes practice and effort.
AI isn’t good for kids, just like social media and sugary drinks. We know this, but the corps making the AI products are trying their very best to get AI into the hands of kids ASAP to avoid regulators and parents saying no. Once it’s entrenched, it’s harder to remove, like social media, or a virus.
(Score: 2) by VLM on Sunday August 31 2025, @04:44PM
TV (at least for xers and boomers, kids don't watch TV anymore) and pop music are also strong analogies.
Probably meant less, in context, but yeah, forced socialization will make people better at socialization, regardless if they like it or not.
(Score: 2) by aafcac on Monday September 01 2025, @05:59PM
It's not, if having 1 child were optimal for parental stress and the development of offspring, we would have gone extinct millennia ago. Couples on average need to have either 2 or 3 children to hit the roughly 2.1 children per reproductive couple rate needed to maintain the population. Any average below 2 results in some sort of a decline and potentially a significant one.
(Score: 3, Interesting) by VLM on Sunday August 31 2025, @04:57PM
The linked essay is kind of cringe because I think the next wave of pro-AI corporate propaganda will be pushing the meme that AI is a new race, so not being a total AI bootlicker in public is the same as being a hateful racist, and in the legacy culture being racist is double plus ungood, so lick them boots, prole.... Of course modern culture is not as narrow minded as legacy culture so thats a whole nother issue where telling them AI-hating is racist is not exactly going to discourage them. To both the bootlickers and the haters, racial jokes including AI are going to unfortunately be a discussion topic, like "An X, a Y, and an AI walk into a bar together and ..." I can already imagine the AI-generated articles in the legacy media like The Atlantic and the NY times.
I don't know how the existing races whom currently rank highly in the oppression olympics will react to being informed in their next NPC download that they've been replaced at the top and now AI has it worse than them. I'm not agreeing with it, but I suspect we're about to get a wave (possibly AI bot generated/posted) of "It's the usual evil white male oppression at work, and that time I used a bulk eraser on a floppy disk back in 1984, what happened to 360K was worse than what happened in Germany in the early 40s to 6M" , etc.
"those people" are so tiresome, but predictable.
(Score: 0) by Anonymous Coward on Saturday August 30 2025, @10:49PM
More boring LLMs are coming and no fun will be allowed for actual adults. They've gotten them to the point were they spit back what you wrote and throw a bow on it instead of formulating replies. By next year, goody2 will be the norm and you'll have to upload IDs to sign up for everything anyways.
The US used to lead the world and now it's citizens are afraid of words from a program. These "kids" are literal teenagers who should have known better and it's just parents trying to get a payday. Had they only put in the effort while they were alive and asking for help....
If that's what we're going to foster from here on out, the country is fucked. One hijacked broadcast telling everyone to an hero oughta be enough to win the next war.