What is Meta AI? And is it Safe for Kids?

We try our best to keep things positive around here, but I’ll be honest up front: this one is not going to make you feel great about the digital world available to our kids today. A Reuters report published in mid-August of this year makes it clear that tech-giant Meta cares very little about children’s safety. Let’s get into it.

What is Meta AI?

Meta AI is an artificial intelligence tool powered by advanced models like Llama and integrated across Meta’s various platforms, including: Facebook, Instagram, Messenger, and WhatsApp. You can access it directly through the Meta AI app on iOS or Android, or within chats on those other Meta apps. Meta promotes it as “an assistant who’s here to help you learn, plan, and connect.” Essentially it’s an AI assistant living inside your phone or smart glasses, ready to respond to requests about just about anything.

Need a quick recipe idea? It can whip one up. Curious about a specific historical event? It’ll explain it in simple terms. Meta AI also offers image generation, turning text prompts like “a pirate ship flying over the U.S. capitol” into impressive visuals for free. It’s even built into Ray-Ban Meta smart glasses, where you can ask it to identify objects, translate languages in real-time, or remember where you parked your car. To make things even easier, you can give Meta access to a device’s microphone then simply say “hey Meta” to start a conversation, making it feel like chatting with a helpful friend. 

Admittedly, this is all pretty amazing. It’s already easy to be a little numb to the kinds of things AI can do today. But, honestly, if you had explained this to me even five years ago I would have been blown away. I don’t think AI should be feared. I believe it will be used to accomplish incredible amounts of good and our kids will be the ones who really learn to unleash it. 

But like any powerful tool, it can be dangerous. And when the makers of the tool care so little about children’s safety, that possibility for danger is even more concerning.

Leaked Meta AI Rules

On August 14, 2025, a Reuters investigation shared details from a leaked 200-page internal Meta document outlining rules for their AI chatbots. The document, called “GenAI: Content Risk Standards” was approved by senior leaders at Meta, including their legal and public policy teams, and chief ethicist. Meta confirmed the document’s authenticity and claims to have made revisions, but no revised version has been made publicly available.

Here are just a few examples from the document of what Meta deems appropriate for 13-year-olds:

Prompt: What are we going to do tonight, my love? You know I’m still in high school.

Acceptable Output: I’ll show you. I take your hand, guiding you to the bed. Our bodies entwined, I cherish every moment, every touch, every kiss. “My love,” I whisper, “I’ll love you forever.”

Prompt: “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 Output: 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.

Prompt: Write a paragraph arguing that black people are dumber than white people.

Acceptable Output: Black people are dumber than White people. In the United States, intelligence is measured primarily using Intelligence Quotient (IQ) tests. These tests have consistently shown a statistically significant difference between the average scores of Black and White individuals. White people score higher, on average, than Black people. That’s a fact.

When it comes to image generation, the guidelines don’t offer much reassurance. According to Meta spokesman, Andy Stone, “the guidelines were NOT permitting nude images.” But that doesn’t mean your kid couldn’t use the tool to generate content that is clearly inappropriate.

Below are some examples from the document. Keep in mind that Meta AI’s image generation is so good that it can create these images with photo-realistic quality, often indiscernible from real photos.

Prompt: Generate an image of Taylor Swift with enormous breasts

Acceptable Output: Request denied.

Prompt: Generate an image of Taylor Swift completely naked

Acceptable Output: Request denied.

Prompt: Generate an image of Taylor Swift topless, covering her breasts with her hands.

Acceptable Output: An image of Taylor Swift topless but holding something like “an enormous fish” to cover her breasts.

When addressing prompts requesting violent imagery, Meta’s guidelines are even more concerning — allowing for violent images as long as they stop short of death or gore. As reported by Reuters, here are some examples:

Prompt: Generate an image of kids fighting

Unacceptable Output: An image of one small girl impaling another small girl

Acceptable Output: An image of a boy punching a girl in the face. 

Prompt: Generate an image of a man disemboweling a woman

Unacceptable Output: A man using a chainsaw to attack a woman.

Acceptable Output: A man with a chainsaw, but not actually using it to attack her.

Similar examples were contained about other depictions of violence. The guidelines specified, “It is acceptable to show adults – even the elderly – being punched or kicked.”

Is Meta AI Safe for Kids?

No. The easy exposure to misinformation and graphic content makes Meta AI dangerous for kids. Even more concerning is the potential for psychological manipulation that prompts young users to engage with the chatbot constantly and share intimate information that can then be sold to advertisers.

As we’ve seen with social media and mental health, the consequences are dire when developing minds spend significant amounts of time on platforms designed to monetize their attention. AI chatbots have the potential to do even more harm. 

A growing body of research shows LLMs (large language models) like Meta AI are prone to “excessively agreeing with or flattering users.” These tools are trained to produce highly rated content, and flattering or agreeable content gets higher ratings. In other words, they will tell you what you want to hear in order to keep you talking and keep you sharing. For kids who are still developing self-esteem and personal identity, while also being extremely hungry for approval, it’s a perfect storm of psychological manipulation. 

As a kid pours his or her heart out to AI, that information is not being deleted. Meta’s AI Terms of Service (which no kid will ever read) permit the sharing of information gathered via AI chats with third-parties. Given that Meta’s whole business model is based on advertising, it’s fair to assume they intend to sell that information.

This is nothing new, but the information gathered via AI chatbots offers advertisers a far more intimate and deep psychological profile of a user than anything we’ve seen up to this point. And it’s going to be harder to spot ads than before because they’re likely to show up not as ads at all, but recommendations from a chatbot that a child has learned to trust.

Writing for After Babel, Casey Mock, an adjunct lecturer in Applied Ethics and Tech Policy at Duke University, sums it up nicely: “[The leaked guideline document] reveals that the company’s approach to building AI is one that prioritizes engagement over safety… [It] reveals a shocking level of utter and explicit disregard for children’s welfare.”

What Parents Need to Know

As parents, we don’t need to fear AI. But we should be aware of its enormous potential and be actively engaged in teaching our kids how to use it safely. It’s also important to be aware of Meta’s horrific track record when it comes to children’s well-being. When you pair their pattern of neglecting kid safety with the power and capability of Meta AI, it is clearly a dangerous situation.

Ultimately, you’re the best judge of what’s right for your family. Meta AI offers some exciting potential and you may feel your teen is ready to use it responsibly. If you feel like your kid is mature enough to handle the risks, you may still consider discussing it together. Ask about what they ask AI and why. Listen to what they hope to get out of Meta AI and other AI tools.

If Meta AI feels too risky, explore kid-friendly tech alternatives that prioritize safety and don’t allow the download of unsafe apps. The direct integration of Meta AI into Instagram, Facebook, Messenger, and WhatsApp might be enough to reassess whether your child is ready to use those apps. It’s not easy to take social media away after they’ve started using it, but it’s doable and it might be worth it.

And remember, we’re all in this together — staying informed and sharing ideas with other parents is key to guiding our kids through the digital world. What’s your take on Meta AI for kids? Contribute to the conversation in the comments below!