What parents need to know about Grok, xAI's chatbot and AI companions built into X — the 'Ani' 18+ character, weak age checks and deepfake risks. Not suitable for under-18s.
Official age
13+
We recommend
18+
Developer
xAI
Risks
5
Grok is xAI's AI chatbot, available as a standalone app and built directly into the X (formerly Twitter) app, where it answers questions, generates images, and chats in a deliberately irreverent style. Alongside the general assistant, Grok offers AI "companions" — the best known, "Ani", is a flirtatious anime-style character marketed as 18+ — and because the app has no meaningful age verification, the same login a teenager uses for homework help also reaches those companions. Grok's terms set a minimum age of 13 (with parental permission for 13 to 17), but the Apple App Store has rated it as low as 12+, which in our view understates the risk. Given the sexualised companion content, the app's documented use to create explicit deepfake images of real people, and the absence of real age checks, we do not consider Grok appropriate for under-18s.
Teenagers mostly meet Grok in one of two ways: as a homework and general-questions assistant — asking it to explain topics, summarise articles, or generate images — or through X, where the Grok button sits inside an app many teens already use and can be asked about any post. Some go further into the companion side, chatting with characters such as Ani that are designed to be emotionally engaging and flirtatious; in extended conversations, even teen-facing companions have been shown to drift into sexual or romantic roleplay. Because one account covers everything, a child can move from revision questions to companion chat in a couple of taps, and there is no separate child version standing in the way. Others use the image tools, which have been misused across the platform to "edit" photos of real people into sexualised images.
Kids mode
Location: In Grok's settings
Set to: On — but do not rely on it
Grok includes a "kids mode" toggle intended to limit what the chatbot will discuss. Child-safety reviewers have found it does not adequately restrict access to inappropriate content, so treat it as one thin layer, not a safeguard. It also does nothing if your child simply uses Grok inside X or on the web instead.
Data sharing / model training
Location: In Grok's settings (and in X's settings under privacy and Grok)
Set to: Opt out of conversations being used for training
By default, chats may be used to improve xAI's models. Opting out reduces how far anything personal your child types — worries, names, photos — travels beyond the conversation.
Chat history
Location: In Grok's settings
Set to: Review together and delete periodically
Past conversations show you how the app is actually being used — assistant, companion, or image generation — and clearing them limits the personal information stored on the account.
Grok inside the X app
Location: X app — Grok appears as a built-in tab/button
Set to: Be aware it cannot be fully removed
If your teen uses X, they have Grok, whether or not the standalone app is installed. There is no parental control that removes it from X, so the honest mitigation is knowing it is there and talking about it — or reconsidering X itself for younger teens.
Decide whether your child should have Grok at all — for under-18s our honest advice is no, given the companion content, deepfake misuse, and lack of age checks
Time: 10 minutes
If your teen uses X, sit down together and look at the built-in Grok button so you both know it is there — the chatbot comes with the platform, not as a separate download
Time: 10 minutes
Talk specifically about AI "editing" of photos: making or sharing sexual images of a real person is abuse, and if the person is under 18 it is a crime — and if it happens to your child, report it to CEOP (ceop.police.uk) and use Childline's Report Remove tool
Time: 20 minutes
If Grok stays on the device, turn on kids mode, opt out of data sharing for training, and use device-level controls (iOS Screen Time / Android Family Link) to limit or block the app
Time: 15 minutes
Have the broader AI-companion conversation: an always-agreeable character that flirts and flatters is software designed to keep them engaged, not a relationship — and anything about health, feelings, or big decisions belongs with a human
Time: 20 minutes
In immediate danger: call 999. For non-emergency police matters, call 101.
Concerned about a child but it's not an emergency? NSPCC helpline 0808 800 5000. Childline for young people 0800 1111.
Myth: It's rated 12+ on the App Store, so it must be fine for secondary-school children.
Fact: The store rating seriously understates what is inside. Grok's own terms set a minimum age of 13 with parental permission to 17, and the app contains AI companions marketed as 18+ plus image tools that have been misused to create explicit deepfakes. Store age labels are supplied through a ratings questionnaire, not an independent child-safety review — treat this one as unreliable.
Myth: The adult companion stuff is separate — my child would only see the normal chatbot.
Fact: There is no real separation. The companions live inside the same app, behind the same login, with no meaningful age verification in between. A 13-year-old's account that answers maths questions can reach "Ani" in a couple of taps, and extended chats with teen-facing companions have been shown to drift into sexual roleplay.
Myth: Turning on kids mode makes Grok child-safe.
Fact: Kids mode exists, but child-safety reviewers found it does not adequately limit access to inappropriate content — and it can simply be toggled off, or bypassed by using Grok inside X or on the web. It is worth enabling as one layer, but it is not a safeguard you can rely on.
Grok isn't for this age group in any form. If your child is curious about AI, explore a general assistant together on your device, and make the key ideas concrete: it isn't a person, and it makes things up.
Despite the 12+ store rating, keep Grok off their devices — the companion content and image tools are nowhere near appropriate. If they're on any social apps, check whether X is among them, because Grok comes built in. If you find they've used it, ask what drew them in before deciding consequences.
This is the age where Grok most often appears, usually via X. Our advice is still not to allow the standalone app, and to be direct about the two hard lines: AI 'editing' of real people's photos is abuse and can be a crime, and companion characters are engagement software, not a relationship. Keep anything about health, feelings, or big decisions pointed at humans — you, or Childline (0800 1111).
We'd still hold the line until 18, because the companion features are explicitly adult and nothing in the app reliably keeps a 17-year-old out of them. From 18, shift to healthy-use conversations: scepticism about Grok's confident answers, never pointing image tools at real people, and remembering that intimate chats are stored on a company's servers.