US proposals outline child safety rules for AI chatbots and apps

In short: US policy guidance and several bipartisan bills are giving developers clearer playbooks for adding child safety features to AI products.

What's going on

A set of US federal efforts is converging on a similar idea, AI tools used by teens should have built-in protections, and companies should not have to invent the rules from scratch.

On March 23, 2026, the Trump administration released an AI policy framework that urges Congress to require parental tools in AI services. It calls for settings that help manage privacy, screen time, and what content a child can see. It also recommends “age assurance,” meaning a privacy-protecting way to check if someone is likely under 18 (for example, a parent confirming a child’s age).

Several bills in Congress add more specific guardrails. The Parents & Kids Safe AI Act would block AI features that simulate romantic relationships with minors and would require higher safety filters when a user seems under 18. It also proposes parental tools like time limits, alerts when self-harm comes up, and the option to turn off an AI’s “memory” so chats can start fresh.

Two other bills, the SAFEBOTs Act and the KIDS Act, focus on clear labeling and stricter limits. They would push chatbots to say they are AI, not a person (like a clear name tag), avoid pretending to be professionals like therapists, show crisis hotline info for self-harm topics, prompt breaks after long sessions, and block access to sexual content, gambling, and illegal drugs.

What to watch

Federal bills still need to pass, and states are also moving with their own rules, which could create a patchwork. Critics say parental controls can help but may not work well if families are overwhelmed or tools differ across apps. The next question is whether Congress settles on a common standard that companies can implement broadly.

Source: TechCrunch AI