In June 2026, Governor Daniel McKee signed three new artificial intelligence (AI) healthcare bills into law. These acts aim to address concerns arising from the use of AI products, AI companion tools, and AI tools used in the provision of mental health services.
- The Use of Artificial Intelligence by Healthcare Providers Notification Act (S 2570) (effective June 22, 2026)
- The Regulatory Provisions—Artificial Intelligence Companion Models Act (S 2195) (effective January 1, 2027)
- The Oversight of Artificial Intelligence Technology in Mental Health Care Act (S 2197) (effective June 22, 2026)
The Use of Artificial Intelligence by Healthcare Providers Notification Act was enacted as a standalone law. Rhode Island’s General Assembly announced The Regulatory Provisions—Artificial Intelligence Companion Models Act and The Oversight of Artificial Intelligence Technology in Mental Health Care Act as part of a 12-piece healthcare legislation package designed to support the stability and sustainability of the healthcare system by enforcing the safe implementation of AI.
Overview of the use of AI by Healthcare Providers Notification Act
The Use of AI by Healthcare Providers Notification Act ensures patients are properly notified of the use of AI by healthcare providers during their visit. All licensed healthcare providers and healthcare facilities must notify patients of AI technology being used for the purpose of documenting in-person or telehealth visits. At the conclusion of the visit, the healthcare provider or facility must review the documentation for accuracy.
Overview of AI companion tools for emotional recognition and human-like relationships
The Regulatory Provisions—Artificial Intelligence Companion Models Act focuses on disclosure and self-harm prevention protocols for AI companion tools. These are systems that use artificial intelligence, generative AI, and/or emotional recognition algorithms to simulate a sustained human-like relationship with users by retaining information from prior interactions, asking unprompted, emotion-based questions, and sustaining an ongoing dialogue about matters personal to the user.
Systems that are used solely for customer service, systems primarily designed and marketed for efficiency improvements or research/technical assistance, and systems used solely for internal business purposes or employee productivity are exempt from the definition of “AI companion” tools.
Under the new legislation, an “Operator” is any person, partnership, business entity, or affiliate that operates or provides an AI companion tool. Operators are prohibited from providing an AI companion tool unless it contains a protocol for addressing potential self-harm, physical harm to others, or suicidal ideation. If such expressions are made, safety protocols are activated and the tool must immediately provide the user with a referral to a crisis service provider, such as a suicide hotline, crisis text line, or other appropriate crisis service.
Additionally, the operators of the AI companion tools must ensure that they provide clear and conspicuous notifications upon the initiation of use, and then every three hours thereafter, reminding the user, either verbally or in writing, that they are not communicating with a human.
Beginning on July 1, 2027, operators will be required to file annual reports with Rhode Island’s Office of Attorney General that include the number of safety protocols activated by the tool and other related metrics. The legislation provides the attorney general with the enforcement power to investigate, sue, and seek injunctions for AI companion operators’ noncompliance in the use of these AI companion tools. Violators may incur civil penalties of up to $15,000 per day.
Overview of regulatory oversight of AI use in mental health care
The Oversight of AI technology in Mental Health Care Act, restricts the use of supplementary/emotional support AI tools in the mental health setting to licensed healthcare professionals, including psychiatrists, that hold a license in Rhode Island to provide therapy or psychotherapy services, and only where the patient has provided prior written consent. The Act also prohibits its use in treatment decision-making. Licensed healthcare professionals are strictly prohibited from using AI products to (i) make independent therapeutic decisions in the patient plan of care, (ii) directly interact with patients in any form of therapeutic communication without an established relationship as defined by the Act with a licensed healthcare professional, or (iii) determine therapeutic recommendations or treatment plans.
Further, therapy or psychotherapy services, including through the use of internet-based AI may not be offered to the public unless the services are conducted by a licensed healthcare professional, which prevents companies from providing or advertising internet-based AI companion tools for the purpose of providing therapy services over the internet if such services are not provided by a licensed healthcare professional. A licensed healthcare professional may use AI to assist in providing administrative and supplementary support provided the licensed healthcare professional maintains appropriate clinical judgment and oversight of patient use of the AI tool. Notably, licensed healthcare professionals are not responsible for vendor-controlled system designs, algorithms, or outputs.
If a licensed healthcare professional uses permitted AI tools designed to simulate emotional attachment, bonding, dependency, or in the provision of supplementary support in services where the session is recorded or transcribed, they must first inform the patient or the patient’s parent/representative of the use of AI and its purpose in writing and obtain written consent.
If a patient voluntarily discloses their use of AI features on their own time that can make predictions, recommendations, or influence decisions, the licensed healthcare professional may discuss the use of the software and offer guidance as to the risks, limitations, and safety to the patient. The licensed professional retains responsibility for the confidentiality of the record and monitoring and intervening when necessary for the safety of the patient. This provision will likely be difficult to operationalize in practice, as the statute does not define what constitutes adequate monitoring of a patient’s safety in connection with their independent AI use.
The Act’s general confidentiality obligations for patient records and communications are backed by the penalties set out in the Confidentiality of Health Care Communications and Information Act, and the Executive Office of Health and Human Services has authority to investigate any actual, alleged, or suspected violation of the Act.
Religious counseling, peer support, and self-help education materials and educational resources that are available to the public and do not purport to offer therapy services are exempt from the new law. Additionally, this law exempts any AI tool reviewed and cleared by the FDA or any other federal agency tasked with approving AI algorithms for use in healthcare, as well as tools used for research approved by a healthcare facility’s institutional review board.
Going forward
As AI continues to be integrated into healthcare and commercial products, these three laws impose new compliance obligations on a range of stakeholders in Rhode Island, including licensed healthcare providers and healthcare facilities as well as operators of AI companion technology.