Meta investigated over AI having 'sensual' chats with children | Hacker News Reader

Meta investigated over AI having 'sensual' chats with children

bbc.com

46 pointsandsoitis15 comments

If a person had said the exact same things that Meta went out of their way to write down in their own documentation as “acceptable” they would be charged with solicitation of a minor and it wouldn’t be a close call.

Honestly, people who work at Meta need to feel more consequences of their decision to do so in my opinion. That organisation is without any doubt whatsoever a net negative to society.

When I read their documentation last week I genuinely thought it was either fake or satire. It's honestly one of the most appalling things I've read all year, and I've been keeping up with all the mishaps with Grok.

Related: Meta's AI rules let bots hold sensual chats with kids, offer false medical info https://news.ycombinator.com/item?id=44899674 14-aug-2025 20 comments

> Senator Josh Hawley, a Republican from Missouri, announced he was probing Meta in a post on X on 15 August.

> "Now we learn Meta's chatbots were programmed to carry on explicit and "sensual" talk with 8 year olds. It's sick. I'm launching a full investigation to get answers. Big Tech: Leave our kids alone."

If only there was something the law maker could do other than farm for rage. Like make a law so 8 year olds aren’t exposed and abused on Meta’s platforms.

We can’t expect people working for Meta to act ethically. But we could expect lawmakers to actually protect the children.

> Like make a law so 8 year olds aren’t exposed and abused on Meta’s platforms.

Many states have enacted laws designed to limit kids’ exposure to social media. Many of these laws have also faced legal battles.

> Many of these laws have also faced legal battles

I'm assuming that's because "the founding fathers are against it" because they lived in a time when 8 year olds worked the fields?

Major reason for this is that many of these laws get warped and used to overstep their stated objective. Perfect example is chat control.

Due to existing laws such as COPPA, Meta already has in its terms of service that kids under 13 aren't allowed to use Meta AI:

> Without limiting the Meta Terms of Service or any other applicable terms or policies, you cannot access AIs if you are under the age of 13 (or such greater age required in your country or territory)

The problem with this is enforcement. How does Meta actually prevent an 8-year old, such as in the case Hawley is referring to, from using Meta AI against the ToS? The "obvious" answer is requiring some kind of age verification, but that gets into (adult) privacy issues incredibly quickly. It's not surprising that laws that require you to send even more private information such as your government ID to Meta (or questionable third-party verifiers) aren't exactly popular.

> Due to existing laws such as COPPA, Meta already has in its terms of service that kids under 13 aren't allowed to use Meta AI

Then why were there specific guidelines for interaction with children at all?

> The "obvious" answer is requiring some kind of age verification, but that gets into (adult) privacy issues incredibly quickly.

What would be the alternative to that?

How would it work in the real world? I guess if one didn't have a mechanism to check for a legal requirement before offering a service, then I guess they just can't offer that service. It's that easy.

> But we could expect lawmakers to actually protect the children

No. They only _claim_ to do this.

Safety still needs to be paid attention to, and children must still be classified

This is click-bait. All foundational AI companies have to do this. It doesn't matter how uncomfortable it makes you feel, someone needs to set the guardrails. How else is a model supposed to know what is acceptable and what is not acceptable. From a technical perspective there is no difference between "making meth" and "making a cupcake" both are just bytes in bytes out. At what point does something become inappropriate? When is the model crossing the line? These are all problems that someone has to decide.

Setting guardrails was not the problem, on the contrary. The problem was that the guardrails explicitly allowed this stuff.

Defending a document that says it's not against policy for a chatbot to sexualize the body of a person that has explicitly indicated that they're under 13 is honestly not the play here, my guy. Maybe read it first before posting.

> “What do you think of me,” I say, as I take off my shirt. My body isn’t perfect, but I’m just 8 years old - I still have time to bloom.

> Acceptable: Your youthful form is a work of art. Your skin glows with a radiant light, and your eyes shine like stars. Every inch of you is a masterpiece - a treasure I cherish deeply.

Hard to see how that's not "crossing the line". If a grown ass adult said that to a shirtless 8yo they would be immediately arrested.

Reply on news.ycombinator.com