As artificial intelligence tools designed to support mental health care spread across the country, states are moving in different directions to regulate them, according to a report by KFF. The result is a patchwork of rules that varies significantly depending on where a person lives, and major questions about safety and effectiveness remain unanswered.
AI mental health tools range from chatbots that offer therapeutic conversation to apps that monitor mood and flag users who may be in crisis. Some are marketed directly to consumers. Others are sold to employers, schools, or health systems. The technology has grown quickly, and in many cases it has outpaced the regulatory systems meant to oversee health products.
The federal government has not established a comprehensive framework for this category of tools. The Food and Drug Administration has authority over some digital health products classified as medical devices, but many AI mental health apps fall outside that definition or exist in gray areas where oversight is unclear. That gap has pushed some states to act on their own.
KFF's reporting found that states are taking meaningfully different approaches. Some have focused on transparency, requiring companies to disclose when a user is interacting with an AI rather than a human. Others have looked at data privacy, given that mental health apps often collect sensitive personal information. A smaller number of states have tried to set clinical standards, raising questions about who should determine whether an AI tool is safe enough to use with someone in psychological distress.
The variation creates challenges for companies building these tools and for the patients using them. A product that meets the rules in one state may not meet the rules in another. Consumers often have little way of knowing what standards, if any, the tool they are using has been held to.
Mental health professionals have raised concerns about tools that simulate therapy without any licensed clinician involved. There are no uniform standards requiring AI mental health products to demonstrate they actually help users or that they do not cause harm. Some researchers have pointed out that people in crisis who turn to an AI chatbot instead of a human counselor could receive inadequate or even dangerous responses.
Supporters of the technology argue that AI tools can expand access to mental health support, particularly in areas where therapists and psychiatrists are in short supply. Wait times for mental health appointments can stretch for months in many parts of the country, and tools that provide immediate, low-cost support have real value for some users.
KFF's report noted that the open questions about how to regulate this space reflect deeper uncertainty about what these tools are, whether they are wellness products, medical devices, or something else entirely. How states and eventually the federal government answer that question will shape how millions of people interact with AI-driven mental health support in the years ahead.
