Four US states moved to restrict AI therapy chatbots in the same week — and the pace is only likely to pick up.

Between April 7 and April 13, 2026, Maine, Nebraska, Missouri, and Maryland all passed legislation that limits how AI can be used in mental health and behavioral health contexts. Each state took a different approach, but the direction of travel is the same: AI cannot substitute for licensed human professionals when it comes to therapy, and any system that implies otherwise faces new legal exposure.

What Happened in Each State

Maine passed LD 2082, which prohibits any person or service from providing, advertising, or offering therapy or psychotherapy unless delivered by a licensed professional. The bill carves out a clear role for AI in the administrative layer: appointment management, billing, and note-taking with client consent are all permitted. What is not permitted is AI acting as the therapist. The bill passed both chambers and was sent to the governor on April 10.

Nebraska passed LB 525, which bundles together the Agricultural Data Privacy Act and the Conversational AI Safety Act. The chatbot portion of the bill sets disclosure requirements: conversational AI services must tell users they are talking to AI, cannot represent themselves as human when a reasonable person would not expect AI, and cannot claim to provide professional mental or behavioral health care.

Missouri included a therapy chatbot ban in an omnibus health care bill. The provisions cover “therapy services, psychotherapy services, or a mental health diagnosis.” The penalty for a first violation is $10,000, enforced by the state’s Attorney General.

Maryland passed HB 148, its focus being AI-driven surveillance pricing and wage manipulation. The bill also includes a provision targeting AI that makes representations or statements relating to behavioral health care without appropriate professional oversight.

The Bigger Picture

These four states are not acting in isolation. Legislators across roughly 30 US states are currently considering more than 60 bills focused on chatbot regulation in 2026. The pattern is consistent: state lawmakers are moving faster than Congress on specific, targeted AI rules, and mental health is one of the clearest battlegrounds.

The federal government’s response has been to push for a 10-year moratorium on state AI laws — but that provision faces significant opposition within the Republican Party, and its fate in the Senate remains uncertain. In the meantime, states are acting.

What this means in practice is a growing patchwork of state-level rules that companies need to track, especially if they operate across multiple states. Tennessee already passed a mental health AI ban with a private right of action. Idaho passed the Conversational AI Safety Act. Washington restricted companion AI chatbots for minors. The list is getting longer every month.

What This Means for Business

If your business deploys AI in any context where employees or customers might discuss emotional or health-related topics, you need to look closely at these laws. The risk is not limited to explicit therapy apps. Customer service chatbots, HR wellness tools, and employee assistance programs could all face scrutiny if they provide responses that sound like counseling without the appropriate disclaimers and professional oversight.

The businesses most exposed are those that:

• Use AI assistants in HR or employee benefits contexts
• Run customer service tools that handle distress, complaints, or sensitive personal matters
• Operate in healthcare or adjacent industries
• Have not updated their AI disclosure language since deploying their tools

A few practical steps are worth taking now. First, audit what your AI systems say when users raise emotional or mental health topics. Second, make sure your chatbots clearly disclose they are AI and direct users to licensed professionals when appropriate. Third, if you are building customer-facing AI tools that might touch these areas, get legal review of the language your system uses.

The regulation here is not trying to eliminate AI from healthcare or wellness contexts. Maine’s law specifically allows AI for administrative support. What these states are drawing is a line between AI as a tool that supports licensed professionals and AI as a replacement for them. That is a line most businesses should be able to operate comfortably behind, as long as they are paying attention.

If you’re building with Claude or Codex right now, grab the free Working With Claude field guide. Thirty-two pages on the full ecosystem, Claude Code in depth, and how to roll agents out properly. Get the free guide.