Rhode Island is not usually the state that shows up first in an AI regulation roundup. This week it is, and the reason is a penalty number specific enough to change a product roadmap: $15,000 a day.
TL;DR: Governor Dan McKee signed three Rhode Island AI laws on June 22, 2026. One bans unlicensed AI therapy services. One requires AI companion products to detect suicidal ideation or self-harm and route users to crisis resources, backed by civil penalties up to $15,000 per day and a private right of action. One requires healthcare providers using AI to document visits to disclose it and review the output. Almost none of this touches a typical small team, unless you build companion or mental-health-adjacent AI, or you're a healthcare provider using an AI scribe.
The three laws, specifically
Therapy chatbot ban. Unlicensed individuals and businesses may not offer therapy or psychotherapy services through AI. The law also bars AI companion products from simulating emotional attachment, meaning products can't be designed to present themselves as having genuine feelings for the user.
AI companion safety protocol. This is the one with teeth. AI companion products must include safety features that detect when a user expresses suicidal ideation, or potential physical or financial harm to themselves or others, and route them to appropriate crisis resources. Noncompliance carries civil penalties up to $15,000 per day, enforced by the state attorney general, with collected fines directed to suicide prevention programs. The law also creates a private right of action: someone physically injured through self-harm, or physically or financially harmed by another person as a result of a violation, can sue directly in superior court for damages and other relief.
Clinical AI-scribe disclosure. Healthcare providers using AI to record or transcribe in-person or telehealth visits must notify patients that AI is being used, and must review the AI-generated documentation for accuracy before it's treated as part of the medical record.
Why the private right of action is the detail worth noticing
Most state AI laws this site has tracked are enforced exclusively by a state attorney general, with no direct path for an individual to sue. Colorado's AI Act is the clean example: no private right of action, AG-only enforcement. Rhode Island's companion-safety law is different. If someone is harmed through self-harm, or harmed by another person because a companion product failed to route a crisis expression appropriately, that person can bring their own lawsuit.
That changes the risk calculation for anyone building or deploying a companion-style product. An AG enforcement action is a risk you can model around resourcing and priorities at one office. A private right of action is a risk you can't predict the timing or volume of, and it scales with how many users your product actually has.
What to actually check
If you don't build or deploy an AI companion, mental-health-adjacent product, or AI therapy service: this is background awareness, not an action item. The obligations fall on providers of those specific product categories.
If you're a healthcare provider, clinic, or practice using an AI scribe or transcription tool for patient visits: confirm your patient notice covers AI use specifically, and confirm someone is actually reviewing AI-generated notes before they're finalized, not just rubber-stamping them. This is the most likely provision to affect a small practice that adopted an AI scribe tool without a formal rollout process.
If you're evaluating or building any product with a companion, emotional-support, or wellness-adjacent framing, even informally: Rhode Island's private right of action means the relevant question isn't just "does our state require this," it's "does any state where our users are located." A companion product used nationally is exposed to the strictest state's rule, not the one where the company is headquartered.
Related Reading
- Multi-State AI Compliance Strategy for 2026
- AI Governance for Healthcare Startups: HIPAA and Beyond
- State AI Law Private Right of Action: Where It Actually Exists
- US State AI Law Tracker (June 2026)
- AI Meeting Transcription and Data Leak Compliance
Sources: Transparency Coalition, "Rhode Island enacts three new AI laws, including a therapy chatbot ban", AI Weekly, "Rhode Island Signs Three AI Laws Covering Mental Health Chatbots", Rhode Island Senate Bill S2195 and House Bill H7350 (2026 session).
