California Assembly Bill 3030, titled "Health Care Services: Artificial Intelligence," requires health facilities, clinics, and physicians' offices that use generative AI to produce written or verbal patient communications about clinical information to include a disclaimer that the communication was AI-generated and instructions for reaching a human. Signed by Governor Gavin Newsom on September 28, 2024, and effective January 1, 2025, it is described as the first US statute imposing AI-disclosure requirements specifically on clinical patient communications.
Status and timeline
AB 3030 was introduced in the 2023–2024 Regular Session of the California Legislature, authored by Assemblymember Lisa Calderon (D-Whittier). Governor Gavin Newsom signed it on September 28, 2024, and it took effect on January 1, 2025. The law is codified as Cal. Health & Safety Code § 1339.75 with amendments to the Business and Professions Code.
Scope and definitions
The law applies to any health facility, clinic, physician's office, or office of a group practice in California that uses generative AI to generate written or verbal patient communications pertaining to patient clinical information. Coverage includes both physician-controlled entities and institutionally licensed facilities.
"Clinical information" is defined broadly to include diagnoses, treatment plans, medication information, lab results, and other medical content. It does not include administrative communications such as scheduling or billing-receipt language, though commentary notes that the boundary is contested in practice because many patient communications blend clinical and administrative content.
Key provisions
When generative AI is used to generate patient communications covering clinical information, the communication must include both a disclaimer that it was generated by generative AI and clear instructions describing how the patient may contact a human healthcare provider, employee, or other appropriate person.
The placement and timing of the disclaimer vary by medium:
- Written communications in physical or digital media (letters, emails, occasional messages): disclaimer prominently at the beginning of each communication.
- Written communications in continuous online interactions (chat-based telehealth, patient-portal chatbots): disclaimer prominently displayed throughout the interaction.
- Audio communications: disclaimer verbally at the start and end of the interaction.
- Video communications: disclaimer prominently displayed throughout the interaction.
The disclaimer is not required if the AI-generated communication is read and reviewed by a human licensed or certified healthcare provider, such as a physician, nurse practitioner, physician assistant, or licensed clinical provider. The exemption is structured to target AI-authored-and-sent communications rather than AI-assisted drafting that still passes through human review.
Enforcement and penalties
Enforcement runs through the existing professional-licensing authority of the Medical Board of California and other relevant licensing boards. AB 3030 establishes no new civil penalty schedule; violations are treated as matters of licensee professional conduct.
Reactions and points of contention
Commentary on the law (including analyses from Duane Morris, Morgan Lewis, Sheppard Mullin, ArentFox Schiff, and Chambers Healthcare AI 2025 – CA) has identified several unsettled questions in applying it.
The "read and reviewed" exemption turns on what counts as meaningful human review. Ambient-scribe tools such as Nuance DAX, Abridge, and Suki present borderline cases in which a physician signs off on AI-drafted content but may not read it closely. The continuous-interaction disclosure requirement for portal chatbots has been described as a potential source of clutter in patient-portal user experience, in tension with usability research on health literacy. The carve-out for billing and administrative communications is contested because patient communications often blend clinical and administrative content.
On enforcement, commentators note that licensing-board enforcement is slow and reactive, and consumer advocates have proposed adding Attorney General enforcement in 2026 amendments. Generative-AI vendors face no direct obligation under AB 3030, though California AI Transparency Act (SB 942) may reach them separately; the combined effect places compliance on healthcare providers rather than vendors. The AB 3030 disclosure does not alter HIPAA minimum-necessary or privacy rules, and vendors still require business associate agreements; some practitioners report redundant layered disclosures in practice. Recent HHS/CMS guidance on AI in clinical use is voluntary (AMA Physician AI Sentiment Report (2026)), and AB 3030 is described as the first binding state clinical-AI communications statute. The American Medical Association has supported transparent clinical-AI use while pressing for clarity on the "human review" standard (AMA Physician AI Sentiment Report (2026)).
Comparison with other disclosure laws
AB 3030 is narrower in scope than the EU AI Act's healthcare-AI regime but more prescriptive on the specific disclosure-medium and timing requirements. The following table situates it among related disclosure regimes tracked in adjacent pages.
| Law | Domain | Mechanism | Exemption |
|---|---|---|---|
| CA AB 3030 | Healthcare clinical communications | Disclaimer + human-contact instructions | Human-reviewed communications |
| Utah Artificial Intelligence Policy Act (SB 149) | Consumer GenAI interactions generally (incl. regulated occupations) | On-request disclosure | Reasonable-person perception test (2025 amendments) |
| California AI Transparency Act (SB 942) | GenAI-provider-side watermarking | Latent + manifest provenance | None |
| California SB 243 — Companion Chatbots | Companion chatbots for minors | Crisis-response + disclosure | None specific |
| California SB 896 (Generative AI Accountability Act) | State-government GenAI use | Citizen-facing disclosure | None |
| EU AI Act (Regulation 2024/1689) Annex III | Healthcare AI (high-risk) | Conformity assessment, data governance, oversight | None |
A closer comparison with the EU AI Act's healthcare provisions shows AB 3030 confined to patient communications and centered on disclosure, while the EU regime extends to the underlying medical-device and decision-support AI:
| Dimension | CA AB 3030 | EU AI Act (Regulation 2024/1689) Healthcare |
|---|---|---|
| Subject | Patient communications | Medical-device + decision-support AI |
| Primary duty | Disclosure | Conformity assessment + CE marking + post-market monitoring |
| Human-review carve-out | Yes | No (limited oversight reduction possible) |
| Penalty | Professional discipline | Up to €35M / 7% turnover |
AB 3030 prompted Medical Board of California guidance (a GenAI Notification resource) and similar guidance from the California Dental Board and Nursing Board. It is part of California's 2024 GenAI legislative package alongside California AI Transparency Act (SB 942) (watermarking), California SB 896 (Generative AI Accountability Act) (state government), and California SB 243 — Companion Chatbots (companion chatbots for minors). Analogous clinical-disclosure provisions, described as less prescriptive, appear in Texas SB 2293 and New York AB 7634.
Related
- related: California AI Transparency Act (SB 942), California SB 896 (Generative AI Accountability Act), California SB 243 — Companion Chatbots, California SB 53 — Transparency in Frontier AI Act — 2024 CA GenAI package
- related: Utah Artificial Intelligence Policy Act (SB 149) — parallel GenAI-disclosure template (consumer-protection frame)
- related: AMA Physician AI Sentiment Report (2026) — physician-profession posture on clinical AI
- related: EU AI Act (Regulation 2024/1689) — healthcare AI is Annex III high-risk
- related: US AI Regulatory Approaches Compared — adds healthcare-clinical dimension
Sources
- Bill text: https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240AB3030
- Medical Board of California GenAI Notification page
- Secondary: Duane Morris (Dec 2024); Morgan Lewis (2024); Sheppard Mullin (Nov 2024); ArentFox Schiff; Chambers Healthcare AI 2025 – CA; CalMatters Digital Democracy