California AI Transparency Act

Summary

The California AI Transparency Act (SB 942, as amended by AB 853 and by SB 1000) has been operative since August 2, 2026. It requires covered providers — any person that creates, codes, or otherwise produces a generative AI system publicly accessible in California — to offer a free disclosure verification tool and to embed latent disclosures in AI-generated or AI-altered image, video, and audio content, with a civil penalty of $5,000 per violation and each day treated as a discrete violation. SB 1000 (Stats. 2026, Ch. 861), an urgency statute signed and effective September 30, 2026, deleted the former 1,000,000-monthly-user threshold from the covered-provider definition.

The California AI Transparency Act is codified at Business and Professions Code Chapter 25 (commencing with Section 22757). SB 942 created the original covered-provider regime in 2024. AB 853 (Stats. 2025, Ch. 674) delayed the operative date and added large-online-platform, GenAI-hosting-platform, and capture-device-manufacturer duties. Two 2026 bills, both approved by the Governor on September 30, 2026, then rewrote parts of the chapter: SB 1000 (Ch. 861) recast the covered-provider duties with immediate effect, and AB 2713 (Ch. 856) rewrote the large-online-platform section before it becomes operative on January 1, 2027. Chapter 25 became operative on August 2, 2026 under Section 22757.6, which neither 2026 bill changed.

SB 942 is one of three California generative-AI transparency statutes and is best read alongside the other two: California AB 2013, the Generative AI Training Data Transparency Act that requires developers to publish training-data documentation, and California SB 53 (TFAIA), which imposes published safety-framework and critical-incident-reporting duties on frontier-model developers.

Compliance date schedule

Chapter 25 does not commence all at once, and the single date most often quoted for it is wrong. SB 942 as enacted in 2024 provided that "this chapter shall become operative on January 1, 2026"; AB 853 amended Section 22757.6 before that date arrived and moved the chapter-wide operative date to August 2, 2026. Three duties AB 853 added carry later operative dates written into the section text itself, so two compliance deadlines under this Act are still ahead. The schedule below is taken section by section from the codified chapter and the chaptered text of SB 1000 and AB 2713 (retrieved October 12, 2026).

What SB 1000 changed on September 30, 2026

SB 1000 (Sen. Becker) passed both houses with its urgency clause, was approved by the Governor and filed with the Secretary of State on September 30, 2026, and took effect that day. It amended Sections 22757.1 through 22757.5 and added Section 22757.4.1:

• No user threshold. A covered provider is now any person that creates, codes, or otherwise produces a GenAI system that is publicly accessible within California (§ 22757.1(e)). The former "over 1,000,000 monthly visitors or users" test is gone, so smaller generative-AI providers are inside the chapter.
• Disclosure verification tool. The "AI detection tool" is renamed a disclosure verification tool. It must let a user assess whether content was created or altered (other than by a minor modification such as cropping, resizing, or color correction) by the provider's system, output detected system provenance data, withhold personal information unless the person concerned consents after a specified notice, accept uploads or URLs, and be callable through an API. A provider may instead direct users to a compliant third-party tool that reads its latent disclosures (§ 22757.2).
• Manifest-disclosure option deleted. The duty to offer users a visible (manifest) disclosure option is removed. The latent disclosure must now also state whether the GenAI system created or altered the content, and from January 1, 2029, whether the system is designed to primarily function as assistive technology (§ 22757.3(a)).
• Licensees. On licensing a GenAI system, the provider must notify the licensee of its obligations. If the provider knows an identifiable licensee modified the system so that it no longer complies, it must, within 72 hours, either terminate the license or notify the licensee, which then has 96 hours to fix or stop using the system and report back. A licensee that does not do so must be reported to the Attorney General (§ 22757.3(b)).
• Assistive technology. Before January 1, 2029, the chapter does not apply to a GenAI system designed to primarily function as assistive technology for individuals with disabilities (§ 22757.5(b)). A false claim to that status carries the separate $50,000 penalty in § 22757.4.1.

Covered-provider duties now operative

As of September 30, 2026, a covered provider must maintain a free, publicly accessible disclosure verification tool, collect user feedback on its efficacy, avoid collecting or retaining personal information from the tool beyond what the chapter strictly requires, and, to the extent technically feasible, embed a permanent or hard-to-remove latent disclosure in image, video, or audio content its system creates or alters. The latent disclosure must convey the provider's name, the system's name and version, the creation or alteration time and date, a unique identifier, and whether the system created or altered the content, and it must be compatible with widely recognized industry standards.

Platform and device phase-in duties

AB 853 adds three later control surfaces. Large online platforms — public-facing social media, file-sharing, mass messaging platforms, and stand-alone search engines that exceeded 2,000,000 unique monthly users in the preceding 12 months — become subject to Section 22757.3.1 on January 1, 2027. AB 2713 (Stats. 2026, Ch. 856), which takes effect the same day, rewrote that section: a platform must detect provenance data embedded into, attached to, or otherwise associated with content; provide a user interface that indicates whether system provenance data or a digital signature identifies content as generated or substantially altered by a GenAI system or captured by a capture device; let users inspect that data on-platform, by link, or by download; and not knowingly strip system provenance data or digital signatures. AB 2713 also provides that the section does not require a platform to maintain, display, or allow download of personal information, or to act on provenance data that is not compliant or interoperable with widely adopted standards. GenAI system hosting platforms may not knowingly make available a GenAI system that does not place the required disclosures; this section also becomes operative January 1, 2027. Capture device manufacturers must provide latent-disclosure options and embed latent disclosures by default for capture devices first produced for sale in California on or after January 1, 2028.

Healthcare and clinical AI routing

SB 942 is a provenance and disclosure statute for AI-generated or AI-altered image, video, and audio content. It should not be treated as the primary clinical-AI privacy or safety rule. Healthcare workflows that handle PHI or ePHI should pair this California content-provenance analysis with the HIPAA compliance for AI in healthcare guide.

Enforcement

A violator of Chapter 25 is liable for a civil penalty of $5,000 per violation, collected in a civil action by the California Attorney General, a city attorney, or a county counsel, and SB 1000 entitles a prevailing plaintiff to reasonable attorney's fees and costs. Each day that a covered provider, large online platform, or capture device manufacturer is in violation is deemed a discrete violation. A civil action brought before SB 1000 took effect cannot be maintained if the alleged conduct does not violate the chapter as amended (§ 22757.4(c)). Falsely representing a GenAI system as primarily assistive technology is penalized separately at $50,000 per violation, each day a discrete violation, until that section is repealed on January 1, 2029 (§ 22757.4.1).

Frequently asked questions

When did the California AI Transparency Act (SB 942) become operative?

The covered-provider duties in Business and Professions Code Chapter 25 became operative on August 2, 2026 under Section 22757.6, and SB 1000 recast them with immediate effect on September 30, 2026. Control surfaces added by AB 853 phase in afterward: large-online-platform and GenAI-system-hosting-platform duties become operative January 1, 2027, and capture-device-manufacturer duties apply to devices first produced for sale in California on or after January 1, 2028.

What is the start date of the California AI Transparency Act?

August 2, 2026, for the chapter as a whole, under Business and Professions Code Section 22757.6. January 1, 2026 is the date SB 942 originally set, and it never took effect: AB 853 (Stats. 2025, Ch. 674) amended Section 22757.6 to August 2, 2026 while SB 942 was still dormant. Two later start dates remain: January 1, 2027 for large online platform and GenAI system hosting platform duties (Sections 22757.3.1 and 22757.3.2), and January 1, 2028 for capture device manufacturer duties (Section 22757.3.3).

Who is a "covered provider" under SB 942?

Since September 30, 2026, a covered provider is any person that creates, codes, or otherwise produces a generative AI system that is publicly accessible within California. SB 1000 (Stats. 2026, Ch. 861) deleted the earlier requirement that the system have over 1,000,000 monthly visitors or users. Before January 1, 2029, the chapter does not apply to a GenAI system designed to primarily function as assistive technology.

What did SB 1000 change in the California AI Transparency Act?

SB 1000 removed the 1,000,000-monthly-user threshold, renamed the AI detection tool a disclosure verification tool, deleted the manifest-disclosure option, required latent disclosures to state whether the system created or altered the content, replaced the licensee contract duty with notice, 72-hour termination-or-notice, and Attorney General reporting steps, exempted primarily assistive-technology GenAI systems until January 1, 2029, and added a $50,000 penalty for falsely claiming that exemption. It took effect on September 30, 2026 as an urgency statute.

What are the penalties for violating the California AI Transparency Act?

A violator of Chapter 25 is liable for a civil penalty of $5,000 per violation, recoverable in a civil action by the California Attorney General, a city attorney, or a county counsel, with attorney's fees for a prevailing plaintiff. Each day a covered provider, large online platform, or capture-device manufacturer remains in violation is treated as a separate, discrete violation, so exposure compounds daily while the statute is operative. A false assistive-technology claim carries a separate $50,000 penalty per violation until January 1, 2029.