Kids and chatbots, what parents should know

Children encounter AI chatbots in more places than a decade ago, including homework help tools, character-based companion apps, in-game assistants, and messaging platforms that have added AI features. The products vary widely in design intent, from tightly scoped tutoring agents to open-ended companions that encourage long conversations. Parents and educators increasingly need a framework for evaluating these tools that goes beyond simple permission checks, because the same underlying technology shows up in many different wrappers.

Children's data protection rules set one of the baselines. In the United States, the Children's Online Privacy Protection Act applies to services directed at children under 13 and has been enforced in AI contexts. In the United Kingdom, the Age Appropriate Design Code, issued by the Information Commissioner's Office, sets expectations for how services used by children handle data, defaults, and nudges. Several EU member states are actively updating guidance under the General Data Protection Regulation and the emerging AI Act.

Product design choices matter as much as legal status. Does the tool clearly tell the child that it is not a person. Does it encourage time-limited sessions or push engagement metrics. Does it have age-appropriate topic filters, and are they transparent to parents. Does it log conversations in a way that parents and schools can review. Answers to these questions rarely appear in marketing material, but they often appear in privacy policies, developer documentation, or independent reviews.

Editor's note: A blanket ban on chatbots is neither realistic nor educationally desirable. Children will meet these tools in school, at friends' homes, and online regardless of household rules. The more useful parental stance is active, informed engagement, trying the tools, discussing what they can and cannot do, and treating chatbot literacy as a skill rather than a threat. Schools that integrate this into existing digital citizenship curricula tend to get better outcomes than those that treat AI as a separate crisis.