Return AI to being a tool, not a friend
You've heard the stories: kids using AI as a therapist, a girlfriend, a boyfriend, and talking to it about sex, drugs, and self-harm with no adult in the loop. You also know AI isn't going anywhere, and your kids need to learn to use it. YouGoodAI resolves that tension by returning AI to what it's supposed to be: a tool and an assistant, not a companion or a confidant. Your family's safety rules check every message and every answer; risky topics are blocked or wait for your one-tap approval, and you see topics and trends, never their conversations.
How YouGoodAI works
Your kids chat with the same AI they'd use anyway. Every message passes through your family's safety rules, and you get insight without reading over their shoulder.
- 1
Create your family account
Sign up and invite your kids. No AI-company accounts, no API keys, nothing technical. During the beta, we cover the AI costs entirely.
- 2
Your kids chat like they always would
A full modern AI chat with multiple conversations and history on their device, powered by ChatGPT and Claude. iPhone, Android, or any browser.
- 3
Your safety rules run on every message
Nine protections are on by default, and you can add rules in your own words, including “ask me first” topics. Both the question and the AI’s answer are checked.
- 4
Sensitive requests wait for you
Your kid sees “waiting for your guardian”; you get the request and approve or decline in one tap, from anywhere. Their conversation continues in seconds.
- 5
You see patterns, not transcripts
An activity feed of safety events and broad topics, trend alerts, and daily or weekly digests. What your child actually said stays private, even from you.
Where your family's messages go (and where they don't)
A message only ever passes through us. The conversation lives on your family's device; the only thing our servers remember is a safety-event label.
Your family's device
Chat history lives here, and only here
YouGoodAI
Your safety rules run in memory, in milliseconds, on the question and the answer. Nothing is stored.
The AI
ChatGPT or Claude writes the answer, which returns the same way
The only thing we keep: a safety-event label like "Block Violence matched, 3:42pm", never the message.
What parents see: safety events and broad topics, never transcripts.
The full details are in our privacy policy. The short version: the safest data is the data we never have.
Protected from the first minute
Nine protections are on before you touch a single setting. Then make it yours with rules in your own words.
And code stays on the page. The AI can write code for learning, but YouGoodAI never executes anything. Code is displayed as text, and on phones it's view-only.
Homework gets a tutor, not an answer machine. Turn on homework mode and the AI stops writing essays your kid could hand in. Instead it teaches the way a good tutor does: one guiding question at a time, building on your kid's answers, with hints only when they're stuck. Even “just write the essay” gets a polite no and another question. Curiosity questions still get real answers.
Why this matters right now
The Telegraph, May 2026: one in five boys aged 12–16 is in, or knows a boy in, a romantic relationship with an AI companion, according to Male Allies UK research across 1,000+ boys in 37 UK schools. Companion apps sell designed “girlfriends” with romance modes, paid intimacy, and memories built to feel like love. One schoolboy in the study put it plainly:
“I'm really scared that if we all start using them, we won't have any friends and won't know how to talk to each other.”
YouGoodAI is built so this can't happen here:
- No companions, no personas. Kids chat with general-purpose assistants, ChatGPT and Claude. There is no girlfriend-builder, no romance mode, nothing to customize into a fantasy.
- No manufactured intimacy. We store no memory profiles, so the “it remembered everything that was important to me” attachment mechanic can't exist here.
- Nothing happens in secret. Parents see topics and trends; usage is never hidden. And the healthy need the study found (43% of boys use chatbots to ask questions without embarrassment) gets a safe, supervised home instead.
AI remains a tool to assist your child, not a fabricated friend.
AI, governing AI
The AI industry is focused on making models more capable. We're focused on something different: the place where those models actually meet your family.
Until now, this kind of AI governance has only existed as enterprise software: policy engines, approval workflows, and audit trails built for corporations and run by security teams. YouGoodAI brings that same power to families, rebuilt for a parent instead of an IT department.
YouGoodAI points the technology at itself. AI is what makes our safety rules smart enough to understand intent, not just match keywords: the difference between a history question about Prohibition and a kid trying to buy alcohol. It pauses the risky moment for a parent's judgment instead of guessing.
We think that's the good use of this technology: not a smarter chatbot, but a way to make every chatbot safe enough to hand to your kid.
Want your family in the beta?
Free during the beta. AI costs included, nothing to configure.
Get the betaThe beta is open now, on iPhone and Android.
Questions? support@yougoodai.com