SB 2037 creates a new framework in Oklahoma law governing the use of artificial intelligence by licensed mental health professionals and licensed health care providers. The bill defines key terms such as artificial intelligence, consent, therapeutic communication, supplementary support, administrative support, therapy or psychotherapy services, and health care services. It requires written notice and affirmative consent before AI may be used to assist in certain therapy, psychotherapy, or health care services, especially where a session is recorded or transcribed.
For mental health services, the bill allows AI to assist only with administrative or limited supplementary support tasks, and only if the licensed professional remains fully responsible for the system’s outputs and data use. It expressly bars AI from making independent therapeutic decisions, directly communicating with clients in therapeutic settings, generating treatment plans without review, or detecting emotions or mental states. The licensed mental health professional must make the final decisions, and violations may result in licensure-board discipline and fines of up to $10,000 per violation. The bill also prohibits offering internet-based AI therapy or psychotherapy to the public unless the service is conducted by a licensed mental health professional, and authorizes the Attorney General to investigate and fine violators.
For broader health care services, the bill imposes similar informed-consent requirements and limits AI from making independent medical decisions, directly interacting with patients in medical communication, diagnosing conditions, or generating advice or treatment plans without review. As with mental health services, the licensed provider must retain final decision-making authority, and violations can trigger disciplinary action and fines up to $10,000 per violation. The bill includes an emergency clause, meaning it would take effect immediately upon passage and approval.
The overall sentiment reflected by the bill’s structure is cautious and protective rather than anti-technology: it permits AI use, but only under strict supervision, disclosure, and consent requirements. There is no recorded vote or committee transcript in the provided materials, so there is no direct evidence of debate or opposition in the available record. Based on the text alone, the bill appears aimed at reassuring patients and preserving professional accountability while allowing limited AI-assisted administrative and support functions.
The main point of contention likely concerns the scope of AI use in clinical settings, especially whether the bill is too restrictive for innovation or sufficiently protective of patients. The strongest restrictions are on direct patient interaction, diagnosis, therapeutic recommendations, and emotion detection, and the bill also targets internet-based AI therapy services unless a licensed professional is involved. Those provisions would most directly affect AI vendors, telehealth platforms, mental health practices, and health care providers using automated tools.
SB 2037 would add new provisions to Title 63 of the Oklahoma Statutes regulating AI use in therapy, psychotherapy, and health care delivery. It would require written disclosure and informed consent before AI is used in covered services, prohibit certain autonomous AI functions, preserve final decision-making authority for licensed professionals, and create disciplinary and fine authority for licensure boards and the Attorney General. The bill would directly affect licensed mental health professionals, licensed health care providers, AI vendors serving clinical markets, and entities offering internet-based therapy or psychotherapy services in Oklahoma.
The bill’s tone is precautionary and consumer-protective, emphasizing informed consent, professional oversight, and limits on autonomous AI in clinical settings. Because no committee transcript or vote record is provided, there is no documented public debate to measure support or opposition. From the text alone, the measure appears designed to balance limited AI assistance with strong safeguards for patients and clients.
The likely points of contention are the bill’s restrictions on AI in clinical decision-making and patient interaction, and its prohibition on internet-based AI therapy unless a licensed mental health professional is involved. Supporters would likely favor the consent requirements, accountability rules, and bans on autonomous diagnosis or therapeutic communication, while critics may argue the bill could hinder innovation, telehealth tools, and administrative efficiency. The Attorney General enforcement authority and $10,000-per-violation penalty provisions may also be debated as potentially broad or punitive.