An act relating to neurological rights and the use of artificial intelligence technology in health and human services

The entire Chapter 233 (AI in Health Care) is absent from the unofficial version. This chapter contained: definitions for generative AI, mental health chatbots, and related terms; a requirement that health care providers disclose when generative AI generates patient communications; enforcement authority with $2,500 per violation penalties; protections for mental health chatbot users' personal data; advertising restrictions on chatbot suppliers; mandatory disclosure that chatbots are AI; an affirmative defense framework for chatbot suppliers who file compliance policies; and detailed restrictions on AI use in health insurance utilization review requiring human providers to make medical necessity determinations.

Sections 1892 through 1895 of the Neurological Rights chapter are absent from the unofficial version. These included: definitions for brain-computer interface, neural data, neurotechnology, consciousness bypass, and written informed consent; a prohibition on collecting or sharing neural data without written informed consent; revocation-of-consent procedures with a 10-day data destruction requirement; requirements for specific written consent before a brain-computer interface can bypass conscious decision making; and enforcement provisions treating violations as unfair or deceptive acts under the Consumer Protection Act with civil penalties up to $10,000 per violation.

In the AI Advisory Council membership, the unofficial version combines the Commissioner of Health and Secretary of Human Services into a single seat (item H: 'the Commissioner of Health Secretary of Human Services or designee'), whereas the official version lists them as separate members (Commissioner of Health in item H, Secretary of Human Services in item J).

The effective date was changed from 'July 1, 2026' in the official version to 'on passage' in the unofficial version, meaning the bill would take effect immediately upon gubernatorial signature rather than on a fixed future date.

Section numbering was renumbered throughout the unofficial version to account for the omitted sections (e.g., the report section is Sec. 4 instead of Sec. 5, and the effective date is Sec. 5 instead of Sec. 6). The cross-reference to the sunset provision was updated from 'Sec. 4' to 'Sec. 3' accordingly.

Added a full definitions section (Section 1892) for the neurological rights chapter, defining brain-computer interface, conscious decision making, consciousness bypass, neural data, neurotechnology, and written informed consent.

Added an entirely new Chapter 233 (AI in Health Care) with three subchapters. Subchapter 1 defines key terms including generative AI, mental health chatbot, mental health provider, and Vermont user, and requires health care providers to include a disclaimer when generative AI generates patient communications about clinical information, plus instructions for contacting a human provider. Exception applies if a licensed human provider reads and reviews the communication.

The AI Advisory Council sunset date was extended from June 30, 2027 to June 30, 2030 (shown as '2027 2030' indicating the amendment).

Added Section 1893 (Neural Data Privacy): prohibits collecting or recording neural data from a brain-computer interface without written notice and informed consent; prohibits sharing with third parties without written request identifying the third party and purpose, plus informed consent; grants right to revoke consent at any time; requires destruction of records within 10 days of revocation and immediate notification to all third parties.

Added Section 1894 (Consciousness Bypass Limitations): prohibits allowing a brain-computer interface to bypass conscious decision making without specific written informed consent for each category of action; provides that consent obtained via consciousness bypass is not valid informed consent; grants right to revoke, including by an agent, guardian, or surrogate.

Added Subchapter 2 (Mental Health Chatbots): Section 9761 prohibits suppliers from selling or sharing individually identifiable health information or user input with third parties (with narrow exceptions for provider requests, health plan requests, and functional sharing under HIPAA-like rules). Section 9762 restricts advertising within chatbot conversations and prohibits using user input to target or customize ads. Section 9763 requires clear disclosure that the chatbot is AI, not a human, before first use, at the start of interactions after 7 days of inactivity, and whenever asked.

Added Subchapter 3 (Health Insurance): Section 9771 requires health plans using AI for utilization review to ensure the tool bases determinations on individual clinical history and circumstances (not solely group datasets), complies with state and federal law, does not supplant provider decision-making, does not discriminate, is open to audit by the Department of Financial Regulation, and does not cause harm. Critically, subsection (b) prohibits AI from denying, delaying, or modifying health care services based on medical necessity - that determination must be made only by a licensed human health care provider.

The report requirement (Section 5) was restructured: the phrase 'in consultation with interested stakeholders' was removed from subsection (a), and the specific sub-items under the report recommendation (neurological rights protections, genAI guidance for regulated professions, utilization review regulation) were simplified in the main text but retained in the clean summary version.

Added Section 1895 (Penalty and Enforcement): violations of the neurological rights chapter constitute unfair or deceptive acts under Vermont's Consumer Protection Act; civil penalty up to $10,000 per violation; Attorney General has full investigative and enforcement authority; consumers have the same rights and remedies as under the Consumer Protection Act.

Added Section 9753: Attorney General may impose administrative penalty of up to $2,500 per violation of the AI in Health Care chapter and may file actions in Superior Court with full Consumer Protection Act authority.

Added Section 9764 (Affirmative Defense): a mental health chatbot supplier can raise an affirmative defense to claims of unlawful or unprofessional conduct if it has a written policy meeting detailed safety, accuracy, fairness, and privacy requirements, filed with the Attorney General for a $100 fee, and complied with that policy. The policy must include clinical best practices alignment, testing, harm identification, real-time acute risk response protocols, user safety prioritized over engagement metrics, anti-discrimination measures, and HIPAA compliance.

Effective date changed from 'on passage' (unofficial version) to July 1, 2026 in the main statutory text of the official version. Note: the clean summary section at the end of the official document still reads 'on passage,' creating an internal inconsistency.

In the neurological rights section, item (3) appears as 'cognitive liberty' in the main statutory text of the official version, compared to 'nondiscrimination in the development and application of neurotechnologies' in the unofficial version. However, the clean summary at the end of the official document still shows the 'nondiscrimination' language, suggesting a possible drafting inconsistency.

The neurological rights list in Section 1891 was changed: 'cognitive liberty' (House version item 3) was replaced with 'nondiscrimination in the development and application of neurotechnologies' in the enacted version.

AI Advisory Council membership was reorganized: the Commissioner of Health position was replaced with the Secretary of Human Services, the public education member (appointed by Vermont-NEA) was given its own distinct position rather than being combined with another, and a new position was added for a member with relevant knowledge and expertise appointed by the Governor.

The effective date was changed from July 1, 2026 to 'on passage,' meaning the law took effect immediately upon the Governor's signature on May 18, 2026, roughly six weeks earlier than originally scheduled.

The required report to the General Assembly was expanded: item (A) now includes 'proposed definitions for relevant terminology' in addition to protections for neurological rights and neurotechnologies.

Sections were renumbered: the report section moved from Sec. 5 to Sec. 4, and the effective date section moved from Sec. 6 to Sec. 5, with internal cross-references updated accordingly.

Entire Chapter 233 on AI in Health Care was removed, including: notice requirements for generative AI patient communications (Sec. 9752), all mental health chatbot regulations including data protection, advertising restrictions, disclosure requirements, and affirmative defense provisions (Secs. 9761-9764), and all restrictions on AI use in health insurance utilization review (Sec. 9771).

AI Advisory Council membership was expanded: a new member with experience in public education appointed by Vermont-National Education Association was added, a new member with relevant knowledge and expertise appointed by the Governor was added, and the Commissioner of Health seat was changed to Secretary of Human Services.

All enforcement mechanisms were removed: the $10,000 civil penalty for neurological rights violations, the $2,500 administrative penalty for AI in health care violations, and Attorney General enforcement authority under both chapters.

Neural data privacy rules (Sec. 1893), consciousness bypass limitations (Sec. 1894), and all definitions (Sec. 1892) from the neurological rights chapter were removed, leaving only a statement of individual rights with no operational rules.

The reporting requirements in Sec. 4 were expanded: the Council must now act 'in consultation with interested stakeholders,' and the report must specifically address neurological rights protections, guidance on generative AI by regulated professions, and regulating AI in health insurance utilization review.

A new neurological right was added: 'nondiscrimination in the development and application of neurotechnologies' as item (3) in Sec. 1891.

Effective date changed from July 1, 2026 to immediate upon passage.