Why I built an AI companion designed for kids instead of letting them use ChatGPT alone
A few months ago I read about kids forming deep emotional attachments to AI chatbots, some to the point of real distress when the app changed or went down, because at that age they genuinely can't tell the difference between a friend and a very convincing pattern of words. That stuck with me. Kids are already meeting AI today, on tools built for adults, with no framework, no supervision, and nobody explaining to them what they're actually talking to.
I didn't want that to be my own kid's first experience with AI. So I built Yoggi, a conversational AI companion designed specifically for children aged 3 to 15, with parents in the loop rather than locked out. It always identifies itself as an AI when asked. Every message is screened in real time, and if something concerning comes up, the parent gets an alert by email immediately, not days later. There's a daily message limit set by the parent, and a nightly summary of what the child talked about and how they seemed.
The goal isn't to give kids a smarter toy. It's to make sure their first contact with AI happens somewhere safe, age appropriate, and supervised, instead of alone on a platform that was never built with them in mind.
Curious how other builders working with kids or sensitive audiences think about this. Where do you draw the line between giving a child a useful tool and protecting them from getting emotionally hooked on it?
I like that so many of your product decisions optimize for a healthy relationship with AI instead of simply maximizing engagement.
I'm curious what you've learned from parents that most changed your thinking. Was there a moment where feedback caused you to prioritize trust and boundaries over something that would likely have increased usage?
That’s exactly the trade-off I’ve been thinking about.
One recurring concern from parents wasn’t only inappropriate content, but the risk of children developing an unhealthy attachment to AI or simply spending too much time with it. That pushed me to make choices that may objectively reduce retention: parent-controlled daily limits, no language designed to create emotional dependency, and clear redirection toward trusted adults when a child needs real human support.
I could probably increase usage by making Yoggi more persistent, more emotionally rewarding, or harder to leave. But for a children’s product, trust matters more to me than maximising engagement. I’d rather parents feel comfortable with their child using Yoggi for 15 healthy minutes than optimise for hours of daily use.
Appreciate the context.
Would be good to continue the conversation as you learn how parents respond to those choices.
What's the best email to reach you on?