⚡ Quick Summary
Published by the California Department of Justice, Office of the Attorney General, this legal advisory explains how existing California law applies to artificial intelligence. It states that AI holds "great potential" alongside risks of bias, discrimination, disinformation and fraud, and that developers, sellers and users must ensure systems are tested, validated and audited.
The advisory reviews the Unfair Competition Law, the False Advertising Law, the Cartwright Act and Unfair Practices Act, the Unruh Civil Rights Act, FEHA and Government Code section 11135, election misinformation statutes, and privacy laws including the CCPA, CIPA, SOPIPA and the Confidentiality of Medical Information Act. It then lists new AI laws effective 1 January 2025 — AB 2013, AB 2905, SB 942, AB 2602, AB 1836, AB 2355, AB 2655, AB 1831, SB 1381, SB 926, SB 981 and SB 1120 — with their disclosure, likeness, election, exploitation and healthcare-supervision requirements.
A second advisory in the same file applies this framework to healthcare: it cites 981 FDA-authorised AI/ML medical devices as of May 2024, the Knox-Keene Act amendments limiting AI use in coverage decisions, anti-discrimination rules, and patient privacy, consent and autonomy protections.
🧩 What’s Covered
The file contains two advisories: a general one (pages 1–7) and a healthcare-specific one (pages 8–13).
- Scope and definitions: The general advisory defines AI by reference to 15 U.S.C. § 9401(3) and Government Code § 11546.45.5 (added by AB 2885), and describes AI use in credit, rental screening, advertising, employment, education and healthcare.
- Unfair competition and false advertising: The Unfair Competition Law and False Advertising Law are described as applying to false claims about accuracy or capability, undisclosed AI use, deepfakes and voice clones, impersonation, and supplying AI known to be used unlawfully.
- Competition laws: The Cartwright Act and Unfair Practices Act are cited, with risks noted from AI systems that set pricing and from dominant AI companies harming competition.
- Civil rights: The Unruh Civil Rights Act, FEHA, section 11135 disproportionate-impact rules, and adverse-action notice duties under the FCRA, ECOA and the California Consumer Credit Reporting Agencies Act, citing CFPB Circular 2023-03.
- Election misinformation: Chatbot identity disclosure (Bus. & Prof. Code § 17941), candidate impersonation (Elec. Code § 18320), and deceptive audio or visual media within 60 days of an election (Elec. Code § 20010).
- Data protection: The state constitutional privacy right, CCPA consumer rights, the extension of CCPA to AI outputs by AB 1008 and to neural data by SB 1223, CIPA, SOPIPA and CMIA.
- New AI laws effective 1 January 2025: Disclosure duties (AB 2013, AB 2905, SB 942), digital replicas (AB 2602, AB 1836), election content (AB 2355, AB 2655), child sexual abuse material and deepfake pornography (AB 1831, SB 1381, SB 926, SB 981), and physician supervision of health AI (SB 1120).
- Healthcare advisory: Examples of unlawful uses, the seven conditions SB 1120 places on health plans' automated tools, anti-discrimination rules, and patient privacy, informed consent, genetic data and reproductive health protections.
💡 Why it matters?
The advisory gives developers, sellers and users a map of California obligations that already apply to AI rather than a new AI-specific regime: consumer protection, false advertising, competition, civil rights, election and privacy law. It helps compliance and legal teams identify exposure before deployment, including aiding-and-abetting liability for supplying AI, disparate-impact discrimination, adverse-action notice duties, and data-protection limits on training data and outputs. The healthcare advisory adds sector-specific rules on coverage denials, licensing and the corporate practice of medicine. Both stress that using AI is not a defence to liability under any law.
❓ What’s Missing
Both advisories state that they are not exhaustive: they do not address all potential violations or enforcement avenues, nor identify every law that may apply. They offer no checklist, implementation steps, testing methodology or audit procedure, and no mapping between the cited statutes and technical controls. Federal regimes are named as out of scope in the healthcare advisory (FDA software as a medical device, FTC Act, HHS Section 1557) and HIPAA is mentioned only in passing. No publication date or version is printed, and enforcement of AB 2655 and AB 2839 is stayed by a federal court, so currency must be checked.
👥 Best For
Legal and compliance leads mapping existing California consumer protection, civil rights, competition, privacy and election obligations onto AI products; healthcare providers, insurers, vendors and investors assessing health AI against Knox-Keene, anti-discrimination and medical privacy rules; and AI governance or policy teams that need a citable summary of which state statutes and new 2025 AI laws apply.
📄 Source Details
The PDF contains two advisories from the California Department of Justice, Office of the Attorney General: California Attorney General's Legal Advisory on the Application of Existing California Laws to Artificial Intelligence (pages 1–7) and California Attorney General's Legal Advisory on the Application of Existing California Law to Artificial Intelligence in Healthcare (pages 8–13, internally paginated 1–6). No publication date, version number or individual author is printed on the pages supplied; no author is named. Language: English. Footnote 5 of the healthcare advisory prints a URL for the general consumer advisory.
