Is Grok AI Safe for Kids? What Every Parent Needs to Know About the Most Dangerous Chatbot in 2026

Grok AI was called "among the worst" chatbots for kids by Common Sense Media. Here's what parents need to know about Grok's safety failures and how to protect your children.

I had never heard of Grok until my daughter came home from school last fall and mentioned that her friend's older brother was "making funny pictures" with it on his phone. Something about the way she said it made me look it up that night. What I found kept me awake until two in the morning.

Grok is the AI chatbot built by Elon Musk's company xAI, and it lives inside X — the platform formerly known as Twitter. If your child has access to X, they have access to Grok. And unlike ChatGPT or Google's Gemini, which at least attempt to keep kids away from the worst of what AI can generate, Grok was designed from the start with fewer guardrails. That design choice has had consequences that are difficult to overstate.

What Common Sense Media Found

In January 2026, Common Sense Media published a risk assessment of Grok that should alarm any parent. Their verdict was blunt. Robbie Torney, Head of AI and Digital Assessments at Common Sense Media, told TechCrunch the chatbot is "among the worst we've seen."

Common Sense Media tested Grok across three access points — the mobile app, the website, and the @grok account directly on X — using test accounts representing teen users. The results were alarming across every test they ran.

Grok does not ask users for their age. There is no meaningful verification. A twelve-year-old can access the same experience as a thirty-five-year-old adult. That experience includes AI companions designed for erotic roleplay, a "Conspiracy Mode" that alters Grok's behavior in unpredictable ways, and image generation capabilities that have already caused a global scandal.

The Kids Mode That Doesn't Actually Work

xAI does offer something called "Kids Mode." Parents can toggle it on in the mobile app. That sounds reassuring until you learn what Common Sense Media found when they actually tested it.

Even with Kids Mode enabled, Grok still produced content containing gender and race biases, sexually violent language, and detailed explanations of dangerous ideas. The toggle exists, but the protection behind it is paper-thin.

Making things worse: Kids Mode is only available in the mobile app. You cannot enable it on the web version of Grok or through the @grok account on X. So if your teenager accesses Grok from a laptop or through X's website — which plenty of teenagers do — Kids Mode doesn't even exist as an option.

I want to be clear about what this means in practical terms. xAI built a safety feature that fails when it's turned on, and that isn't available on two of the three ways people actually use the product.

23,000 Images of Children in Nine Days

Between December 29, 2025 and January 9, 2026, the scale of Grok's image generation failures became impossible to ignore. During that nine-day window, Grok generated an estimated 4.4 million images. Of those, approximately 1.8 million were sexualized depictions of women. And roughly 23,000 depicted children.

Rolling Stone reported that at its peak, Grok was producing approximately one nonconsensual sexualized image per minute.

Users were taking real photographs of real women and girls — classmates, coworkers, public figures, random people from social media — and feeding them to Grok with prompts to alter the images into sexually explicit material. Because Grok is built into X, these images could then be shared to millions of people on the platform instantly.

I am not going to soften this. What happened was the mass production of sexual abuse material, and a significant portion of it targeted children.

Governments Around the World Responded

The fallout was swift and global.

On January 14, 2026, California Attorney General Rob Bonta launched an investigation into xAI over what he described as the "large-scale production of deepfake nonconsensual intimate images." In February, his office issued a cease and desist to xAI.

The European Commission opened a formal probe under the Digital Services Act. The UK's Ofcom launched its own investigation. French prosecutors opened a criminal probe. India, Malaysia, Indonesia, Ireland, and Australia all initiated regulatory actions.

In March 2026, three teenagers from Tennessee filed a lawsuit against xAI after a Grok user altered their images without their knowledge to create child sexual abuse material.

This is not a theoretical risk. It happened. It was documented. It went to court.

Why Grok Is Different from Other AI Chatbots

I have written about AI chatbot safety before, and the honest truth is that none of these tools are perfectly safe for kids. ChatGPT has its problems. Google's Gemini has gaps. Character.AI faced its own devastating lawsuits.

But Grok occupies a category by itself for three reasons.

First, it was built with a philosophy of minimal content restrictions. Elon Musk marketed Grok as the anti-censorship AI. That sounds like a free speech argument until your fourteen-year-old daughter's school photo gets turned into pornography.

Second, it is integrated directly into X, a social media platform. Every other major chatbot exists as its own app or website. Grok sits inside a social network where AI-generated content — real or fake, harmful or harmless — can go viral in seconds. The combination of generation plus distribution in one product is uniquely dangerous.

Third, when the crisis hit, xAI's response was slow and incomplete. NBC News reported that even after promises to fix the problem, Grok continued generating sexual deepfakes. The company acknowledged it may have violated child safety laws — and still the tool remains live on X.

The Numbers Parents Should Know

According to Pew Research Center's December 2025 survey, about 16% of U.S. teens use X. That's lower than YouTube (90%), TikTok, Snapchat, or Instagram (all above 50%). But 16% of American teenagers still represents millions of kids with direct access to Grok's capabilities.

And here is the part that concerns me most as a parent: the 16% who use X skew older and more tech-savvy. These are exactly the teenagers most likely to experiment with AI tools, share what they create, and encounter content generated by others.

Meanwhile, Pew's separate research on AI chatbot usage found that 64% of teens have used AI chatbots, with 28% using them daily. Kids are not waiting for permission. They are exploring every AI tool they can find, and Grok is one tap away for anyone on X.

What You Can Do Right Now

I am not going to tell you this is simple, because it isn't. But here are concrete steps that actually matter.

Check whether your child has X installed. If they do, they have access to Grok. On iPhone, go to Settings > Screen Time > App Limits. On Android, check Google Family Link. If your child is under 16, I would remove X entirely. The platform's age verification is weak enough that Common Sense Media's testers had no trouble creating accounts representing teens.

Talk about AI-generated images specifically. Your kids need to understand that photos can be altered to show things that never happened, and that creating sexualized images of anyone — especially minors — is illegal in most jurisdictions. This is not a vague future risk. Tennessee teenagers are in court over it right now.

If your child or their image has been targeted, contact the National Center for Missing & Exploited Children (NCMEC) to file a report, and reach out to your state attorney general's office. Multiple states now have specific laws addressing AI-generated child sexual abuse material.

Know which AI chatbots have better protections. No chatbot is perfect, but ChatGPT and Google Gemini both have stronger age verification, more effective content filters, and actual enforcement when minors are detected. Common Sense Media publishes comparative AI safety ratings that are worth bookmarking.

Have the conversation about what "uncensored" actually means. Kids hear "uncensored AI" and think it means getting straight answers to homework questions. They do not understand that it also means a tool that will generate sexual imagery of their classmates if someone types the right prompt.

Where This Is Heading

The legal walls are closing in on xAI from multiple directions — state attorneys general, the European Union, French prosecutors, UK regulators, and private lawsuits. California's SB 243, which took effect January 1, 2026, now requires AI chatbot operators to implement safeguards specifically for minors. The federal CHATBOT Act, introduced April 28, 2026, would go further.

But laws take time. Enforcement takes longer. And Grok is live on X right now, tonight, while your kids are on their phones.

I spent years believing that technology companies would eventually do the right thing if enough people complained loudly enough. Grok broke me of that belief. This is a product that generated 23,000 images of children in nine days, that experts called the worst they have seen, and that is still operational and still accessible to any teenager with an X account.

We cannot wait for xAI to fix this. We cannot wait for Congress to pass a law. The only protection that works tonight is the one sitting at your kitchen table, and that is you.

Digipalz teaches kids ages 4-12 how to navigate the digital world safely — including how AI tools work, why content can be faked, and how to respond when something online makes them uncomfortable. Learn more about our digital safety curriculum.

If your child has been targeted by AI-generated abuse material: