Why These Three Laws Matter for Multistate Deployments
Organizations deploying AI mental health tools across state lines face a fragmented regulatory landscape. Nevada, Illinois, and Utah have each enacted laws that govern how artificial intelligence may be used in mental health contexts, but they differ substantially in scope, permitted use, and required safeguards. A compliance approach designed for one state may create legal exposure in another.
Statutory Comparison at a Glance
The table below summarizes how the three statutes differ in their core regulatory posture.
| Dimension | Nevada AB 406 | Illinois HB 1806 | Utah SB 226 |
|---|---|---|---|
| Regulatory approach | Prohibition | Restricted supplementary use | Conditional permission |
| Permitted AI role | None in therapy delivery | Administrative support only, under licensed oversight | Direct interaction, with disclosure |
| Core requirement | No claim of providing therapy or acting as a professional | Licensed professional oversight of AI outputs | Clear AI identification and data-privacy safeguards |
Scope Differences: What Each Law Actually Regulates
Nevada's statute focuses narrowly on preventing AI systems from presenting themselves as capable of delivering mental health therapy or holding themselves out as licensed professionals. Illinois takes a broader supervisory approach, permitting AI to assist with administrative and supplementary functions but requiring that any substantive interaction remain under the oversight of a licensed professional. Utah's law is the most permissive of the three, allowing AI chatbots to interact directly with users, provided the tool meets specific disclosure and data-privacy conditions.
Where the Laws Overlap
Despite their differences, all three statutes share an underlying concern: users should not be misled about whether they are interacting with a licensed human professional or an AI system. This shared concern makes disclosure and transparency a common thread across all three jurisdictions, even though the specific mechanisms differ.
Disclosure, Oversight, and Recordkeeping Obligations
Compliance obligations vary in both form and rigor across the three states:
- Nevada requires that AI tools avoid any representation, direct or implied, that they provide therapy or are licensed mental health professionals.
- Illinois requires that licensed professionals maintain oversight of AI-assisted functions and that AI remain limited to administrative or supplementary roles.
- Utah requires clear, conspicuous disclosure that a user is interacting with an AI system, along with data-privacy protections governing how mental health information is collected and used.
Practical implication
A single disclosure banner or a single oversight workflow will not satisfy all three statutes simultaneously. Policies must be configured per jurisdiction rather than applied uniformly.
Enforcement Mechanisms and Practical Risk
Each state's enforcement mechanism reflects its regulatory philosophy. Nevada's prohibition-based model creates risk primarily around misrepresentation, meaning organizations face exposure if their AI tool is perceived as claiming therapeutic capability. Illinois's oversight-based model creates risk around inadequate professional supervision of AI-assisted functions. Utah's disclosure-based model creates risk around insufficient or unclear AI identification and gaps in data-privacy handling.
For organizations operating in more than one of these states, the practical risk is compounded: a configuration that satisfies Utah's disclosure requirements does not automatically satisfy Illinois's oversight requirements, and neither addresses Nevada's outright prohibition on therapeutic claims.
Where Runtime Governance Fits
Because these obligations differ by jurisdiction, compliance leaders need policy enforcement, audit logging, and human oversight controls that can be applied differently depending on where a user is located and which statute governs that interaction. Runtime governance allows organizations to enforce jurisdiction-specific rules, including disclosure language, oversight checkpoints, and prohibited claims, without maintaining entirely separate systems for each state.