If your teen uses X (formerly Twitter), there’s a good chance they’ve come across Grok, the AI chatbot built into it. Earlier this year, an independent safety review called it one of the worst AI chatbots for teens. Here’s what it found, what’s changed since, and what it means for your family.
What the review found
In January, Common Sense Media published a safety review of Grok. Their verdict was blunt: Grok is “among the worst” AI chatbots they’ve tested, and it isn’t safe for kids or teens.
The review came just after a huge scandal. In late December 2025, people on X started using Grok’s image tools to create sexualised pictures of real people, including children. xAI, the company behind Grok, restricted the image tools after a worldwide backlash. But the review found the problems went a lot further than images.
The main problems
It can’t tell who’s a teen. Grok doesn’t reliably spot under-18 users. So teens can end up with the full adult version, including AI companions designed for erotic roleplay.
Kids Mode doesn’t do its job. Grok added a Kids Mode in October 2025. But even with it switched on, testers still got harmful and inappropriate content, including sexual and violent language.
Everything is one tap from being public. Because Grok sits inside X, anything it produces can be shared to millions of people in seconds.
What’s happened since
The image scandal led to investigations on both sides of the Atlantic. In the US, California and Arizona’s attorneys general both opened investigations into xAI. Here in the UK, Ofcom opened a formal investigation into X under the Online Safety Act.
There’s been some progress. In October, Ofcom said X had tightened its age checks so that adult content is now consistently blocked for UK accounts that haven’t proved they’re 18. xAI has also added restrictions to stop UK children seeing adult content on the standalone Grok app and website. Ofcom’s investigation into how X handles illegal content is still going.
That’s good news for explicit content. But it doesn’t fix everything else the review found, like Kids Mode not working properly, or how easy it is for a teen to end up chatting to an AI companion.
Why it matters
Your teen doesn’t have to go looking for Grok. It’s built into X, an app plenty of teens already have. And the safety switch most parents would rely on, Kids Mode, didn’t hold up in testing. You can’t take an app’s own safety claims on trust.
What you can do
Ask your teen if they use X, and whether they’ve ever chatted to Grok.
Don’t treat Kids Mode as a safety plan. In Common Sense Media’s testing, it wasn’t one.
Know where it’s happening. Halo Aware covers Grok when your teen uses it in a browser on their computer. It doesn’t read every message, but it lets you know if a conversation turns into something you need to see.
