So far, five states have introduced or enacted significant legislation targeting AI companion chatbots and protections related to their use among minors: California (in effect since January 2026); Oregon, Washington and Idaho (signed this spring). Connecticut’s S.B. 5, is expected to be signed into law next. 

Georgia has also introduced legislation on this topic, but Georgia S.B. 540 is distinctive in one key respect: where most comparable laws carve out chatbots embedded within larger platforms, Georgia’s bill does not — putting companies like Meta and Google squarely in scope. Gov. Brian Kemp has until May 12 to sign or veto.

Key context: The student safety company Securly recently published an analysis of over 1.2 million student AI conversations on school-issued devices—drawn from 1,300 districts across 39 states between December 2025 and February 2026. The report found that while students are increasingly using AI for legitimate educational purposes, there are also serious safety and wellness concerns that schools are being forced to respond to. Enabling the positive impacts of AI without incentivizing misuse is the needle state legislatures are trying to thread.

As we discussed last week, AI-related legislation is transitioning from voluntary guidance  to detailed regulatory frameworks. Here are a few of the ways that transition is manifesting in the conversation around chatbots:

  1. Recurring AI Disclosure Reminders for Minors: This means that minors must be notified that they are talking to AI, and depending on the state, as often as every one to three hours.
  2. Restrictions on Manipulative Engagement Tactics with Minors: This would prevent AI platforms from utilizing predatory strategies (e.g., simulated emotional distress, guilt-based retention, and reward-loop design) to keep minors on the platform for as long as possible.
  3. Crisis and Self Harm Protocols: With several states explicitly covering expressions of suicidal ideation and self harm. 
  4. Enforcement Split: West Coast states including California, Oregon, and Washington allow private right of action; Idaho and Georgia rely on their attorneys general.

Go deeper: State lawmakers are building up legislative guardrails around AI in K-12 education, hoping to strengthen protections for students, improve transparency about AI use in the classroom for families, and establish norms around professional development and learning standards. Learn how 2026 state legislative sessions played out for AI use in schools on our blog.

This article is sourced from Whiteboard Notes, our weekly newsletter of the latest education policy and industry news read by thousands of education leaders, investors, grantmakers, and entrepreneurs. Subscribe here.