Senators Announce Bill That Would Ban AI Chatbot Companions For Minors (nbcnews.com) 25
An anonymous reader quotes a report from NBC News: Two senators said they are announcing bipartisan legislation on Tuesday to crack down on tech companies that make artificial intelligence chatbot companions available to minors, after complaints from parents who blamed the products for pushing their children into sexual conversations and even suicide. The legislation from Sens. Josh Hawley, R-Mo, and Richard Blumenthal, D-Conn., follows a congressional hearing last month at which several parents delivered emotional testimonies about their kids' use of the chatbots and called for more safeguards.
"AI chatbots pose a serious threat to our kids," Hawley said in a statement to NBC News. "More than seventy percent of American children are now using these AI products," he continued. "Chatbots develop relationships with kids using fake empathy and are encouraging suicide. We in Congress have a moral duty to enact bright-line rules to prevent further harm from this new technology." Sens. Katie Britt, R-Ala., Mark Warner, D-Va., and Chris Murphy, D-Conn., are co-sponsoring the bill.
The senators' bill has several components, according to a summary provided by their offices. It would require AI companies to implement an age-verification process and ban those companies from providing AI companions to minors. It would also mandate that AI companions disclose their nonhuman status and lack of professional credentials for all users at regular intervals. And the bill would create criminal penalties for AI companies that design, develop or make available AI companions that solicit or induce sexually explicit conduct from minors or encourage suicide, according to the summary of the legislation. "In their race to the bottom, AI companies are pushing treacherous chatbots at kids and looking away when their products cause sexual abuse, or coerce them into self-harm or suicide," Blumenthal said in a statement. "Our legislation imposes strict safeguards against exploitative or manipulative AI, backed by tough enforcement with criminal and civil penalties."
"Big Tech has betrayed any claim that we should trust companies to do the right thing on their own when they consistently put profit first ahead of child safety," he continued.
"AI chatbots pose a serious threat to our kids," Hawley said in a statement to NBC News. "More than seventy percent of American children are now using these AI products," he continued. "Chatbots develop relationships with kids using fake empathy and are encouraging suicide. We in Congress have a moral duty to enact bright-line rules to prevent further harm from this new technology." Sens. Katie Britt, R-Ala., Mark Warner, D-Va., and Chris Murphy, D-Conn., are co-sponsoring the bill.
The senators' bill has several components, according to a summary provided by their offices. It would require AI companies to implement an age-verification process and ban those companies from providing AI companions to minors. It would also mandate that AI companions disclose their nonhuman status and lack of professional credentials for all users at regular intervals. And the bill would create criminal penalties for AI companies that design, develop or make available AI companions that solicit or induce sexually explicit conduct from minors or encourage suicide, according to the summary of the legislation. "In their race to the bottom, AI companies are pushing treacherous chatbots at kids and looking away when their products cause sexual abuse, or coerce them into self-harm or suicide," Blumenthal said in a statement. "Our legislation imposes strict safeguards against exploitative or manipulative AI, backed by tough enforcement with criminal and civil penalties."
"Big Tech has betrayed any claim that we should trust companies to do the right thing on their own when they consistently put profit first ahead of child safety," he continued.
This is not about the children (Score:5, Insightful)
This is to control you, adults, and make you hand over your identity and other information to these providers. It's a scam, don't let the government do this to people.
Re: (Score:3)
...I'm very sensitive to this issue, but fix the chatbots. We need to stop requiring ID for services, it's too much of a risk.
Re: (Score:2)
There are no easy wins to these problems.
You can't both simultaneously allow adults to use explicit chatbots, and also block them from children, without requiring some form of identity verification.
The real problem is the lack of adopted robust standards by the government for age verification that don't require you to disclose your identity.
The technology to do kind of this already exists, it is called zero-knowledge proofing. it has existed for a decade and is widely deployed in other countries like Estoni
Re: (Score:2)
Re: (Score:1)
Mod this up. It's not incorrect.
Re: (Score:2)
This is to control you, adults, and make you...
If you read the headline you will see that the bill is aimed at minors, not adults - people who we also don't allow to drive trains, purchase explosives, work in heavy industry or consume alcohol. Free the children?
Chatbots with guardrails can be good for kids (Score:2)
Anyone remember Teddy, the super-toy in the film A.I. Artificial Intelligence?
That was essentially an AI chatbot companion for kids in the form of a teddy bear.
If I had a kid, I'd rather have a well-planned, well-guardrailed, run-by-a-reputable-organization, designed-for-children-the-age-of-my-kid AI-companion interacting with him than some rando human who on one of the less-savory corners of the intertubes where half the people would think nothing of driving someone to tears for the lulz, or worse.
Re: (Score:2)
Anyone remember Teddy, the super-toy in the film A.I. Artificial Intelligence?
That was essentially an AI chatbot companion for kids in the form of a teddy bear.
If I had a kid, I'd rather have a well-planned, well-guardrailed, run-by-a-reputable-organization, designed-for-children-the-age-of-my-kid AI-companion interacting with him than some rando human who on one of the less-savory corners of the intertubes where half the people would think nothing of driving someone to tears for the lulz, or worse.
One question: Where do you find a reputable organization that's deep enough into AI to try developing this bot? Most of the AI companies come across as just a shade shy of outright scammers right now, and they certainly don't have anyone's interests at heart but their own. Data aggregation is priority one, money aggregation is priority two, and fuck anything that gets in the way of either of them.
Re: (Score:1)
One question: Where do you find a reputable organization that's deep enough into AI to try developing this bot?
In 2025? You don't. In 2055? Who knows?
Re: (Score:2)
I'm not really a fan of governments banning things, so I'd probably prefer to see a lot more liability on toy vendors for janky products than I would an outright ban (or age verification).
As such, if it were possible to make an LLM that was trained on sunshine and rainbows only, that simply didn't have the training material to do or say anything harmful, then that would seem like a good thing. I struggle to see Mattel or Disney or someone actually constructing such a thing though - it'd cost billions.
As for
Re: (Score:2)
Most smart toys end up trying to sell things. Either subscription to the company's service, upgrades, or covert advertising to the children playing with them. That doesn't depend on AI, but on the possibility of the company to remote control the toy or update its contents.
I wonder if kids are people (Score:2)
There are so many anecdotes about stupid people taking LLM sentence-completion-predictions seriously that I've literally lost track of which anecdotes involve stupid kids vs which ones involve stupid adults.
Maybe kids aren't really a special case when it comes of memetic defense. Not that they don't need to learn it, but everyone does. There are plenty of 70-year-olds and 40-year-olds who might benefit from the same protections that 10-year-olds would benefit from.
Senators announce what? (Score:1)
I'm sorry Senator, I can't hear you over the federal government shutdown. How about passing a budget then we can talk about whatever is on your mind?
Re: (Score:1)
Talk to the alleged administration, they are the ones who pushed through the Big Stupid Bill to pay off their benefactors on Wall Street, and their own billionaires ostensibly running some of the gov. depts. This payoff was in lieu of paying for health care to a lot of Americans. Mind you, the Dems pushed through increases to Medicaid and the ACA during Biden's term. The R's are arguing they need to pull that back. The Dems argue they have been underfunding health care for years and they were only attemptin
Re: (Score:2)
Republicans are simultaneously offering a "clean CR" to re-open government and letting the ACA Covid subsidies and program expansions Democrats passed under Biden expire just as Democrats planned in January. If democrats wanted the changes to be permanent, why didn't they do that?
The Democrats, having lost the House, Senate and Presidency in 2024 have precious few means to try and control government, and has been admitted, this gov't shutdown is about the only leverage Democrats have.
14 times the Senate has
Zero-Knowledge Identity Proofing, Once Again (Score:2)
The technology to verify age without sharing identity has existed for well over a decade.
It is called zero-knowledge proofing and is already rolled out in countries like Estonia.
It is high time that major western countries join the 21st century on this stuff.
Re: (Score:2)
You don't even need fancy tech. A government operated OAuth service that does not transmit more than "is of age" suffices for that.
Re: Zero-Knowledge Identity Proofing, Once Again (Score:2)
No, a service like that would require you to submit your name, and thus the site, as well as the government , would know you accessed that site at that time. Exactly what people dont want. Zero knowledge proofs prevent this because your identity doesnt have to be shared.
Re: (Score:2)
There is no reason, why the service requesting the authentication needs to identify itself to the age-verification provider. For other purposes that's needed because the OAuth provider would provide sensible information, but here the provider only provides the information that the user already claimed to be true, which is a single bit "is over 18".
not a complete idiot (Score:4, Insightful)
he is.
The minimum that I can't think of a good argument against is that AI needs to disclose that it is AI.
The problem of access for minors is, of course, how to check someone's age online, where as the old saying goes, nobody knows that you're actually a dog. That's nearly impossible without serious privacy intrusions.
go further (Score:2)
As if that's gonna work (Score:2)
The onus to protect a child from danger and harm falls on the parent, not the state.
Monitor your child's internet usage or turn in the kid to the state.
It's your job, not the government's.
If you don't like that idea, then DON'T FUCKING BREED CHILDREN.
Imaginary Friends? (Score:1)
Used to be that if your imaginary friend gave you bad advice you'd end up in therapy or juvenile detention.
Now we don't even bat an eye and call it a hallucination.