Parental controls for AI chatbots in the UK
The short answer
There are three layers, and no single one is enough: the platforms' own controls (real but narrow), device-level screen-time tools (which see the app, not the conversation), and conversation-level monitoring, which is the layer Halo Aware provides.
The honest UK picture: the platforms teens use most have wildly uneven safeguards, and most traditional parental-control apps were built for a pre-AI internet.
Layer 1: the platforms' own controls
ChatGPT has proper linked parental controls (quiet hours, reduced content, feature switches). Google ties Gemini access to Family Link age settings. Character.AI added a parental-insights option. Most of the rest, including the companion apps teenagers actually gravitate to, offer little beyond a self-declared age gate. Where a platform offers controls, turn them on; our safety directory notes what each platform provides, with sources and dates.
Layer 2: device tools, and their blind spot
Screen-time tools (Family Link, Apple Screen Time, and the big parental-control suites) can limit hours or block an app entirely. What none of them can do is tell you what happened inside an AI conversation: two hours on Character.AI could be creative writing or an unhealthy attachment, and a device-level tool cannot tell the difference. They also push a binary choice, block it or go blind, that fits AI badly: these are tools your child will likely use for school and work for the rest of their life.
Layer 3: conversation-level visibility
This is the layer we built. A visible extension in the browser on the computer they use captures AI chats across twelve platforms, including the roleplay apps, and turns them into topic summaries, usage patterns and risk alerts. Raw conversations are processed and discarded; for serious alerts you see two or three lines of context. It's a UK product: prices in pounds including VAT, data under UK GDPR, UK crisis resources in the alerts, and it's currently available in the UK only.
What a sensible UK setup looks like
For most families: platform controls on wherever they exist, ordinary screen-time boundaries you'd have anyway, one house rule that AI chat happens on the computer rather than in phone apps, and summary-level visibility so you learn about patterns while they're still patterns. And all of it with your child's knowledge, which is both the Children's-Code-shaped answer and the one that survives contact with a teenager.
The AI layer, built in the UK
Free for ChatGPT and one child; paid plans from £4.99/month cover all 12 platforms. Prices include VAT, and your child knows it's there.