California AI Laws: SB 53, SB 243, SB 942 and More (2027)

As of November 2026, California has no single comprehensive AI act; it regulates AI by subject, through the Transparency in Frontier Artificial Intelligence Act (SB 53), the California AI Transparency Act (SB 942, rewritten by SB 1000), the companion chatbot law (SB 243), the training-data law AB 2013, health-care AI statutes, and state regulations on automated decisions.

Status as of November 4, 2026: SB 53, SB 243 and the AI Transparency Act are in force. So is AB 2013, on the docket entries we read through October 21, 2026, which show no order blocking it. The largest group of new rules starts January 1, 2027, including CCPA automated-decision duties and platform provenance duties. SB 947 (workplace decisions) and the main duties of SB 1119, "Adam's Law" (children and chatbots), start July 1, 2027. The Ninth Circuit has set argument in xAI's challenge to AB 2013 for November 18, 2026.

This page covers California state AI law. Federal law appears only where it touches a California statute, and election or intimate-image deepfake law has its own California page.

California AI Laws at a Glance

California has no law that regulates AI in general. It regulates by subject. Each date below comes from the law's own text or, where the text states none, from the January 1 general rule explained in the 2027 section.

SB 53: California's Frontier AI Law (TFAIA)

SB 53, the Transparency in Frontier Artificial Intelligence Act, is the California law aimed at the largest AI models. Governor Newsom signed it September 29, 2025 (Ch. 138, Stats. 2025), and it took effect January 1, 2026. As of November 4, 2026, by our count of enacted laws, three states have frontier-model safety statutes. California's is in force. New York's and Illinois's start January 1, 2027.

Who is covered. A "frontier model" is a foundation model "trained using a quantity of computing power greater than 10^26 integer or floating-point operations." A frontier developer is a person who has trained, or begun training, such a model. A "large frontier developer" is one that "together with its affiliates collectively had annual gross revenues in excess of five hundred million dollars ($500,000,000) in the preceding calendar year."

What every frontier developer must do. Before or when deploying a new or substantially modified frontier model, it must publish a transparency report. The report lists the developer's website, a contact mechanism, the release date, languages, output modalities, intended uses, and generally applicable restrictions. A critical safety incident goes to the Office of Emergency Services within 15 days of discovery. If an incident poses an imminent risk of death or serious physical injury, the developer must disclose it within 24 hours to an authority, including a law enforcement or public safety agency with jurisdiction.

What large developers must add. A large frontier developer must publish a frontier AI framework, review it at least once a year, and publish any material change within 30 days. Its transparency reports must also summarize catastrophic-risk assessments, their results, and the extent of third-party evaluators' involvement. SB 53 defines catastrophic risk as a foreseeable and material risk that a frontier model will materially contribute to the death of, or serious injury to, more than 50 people, or more than $1 billion in property damage, from a single incident in which the model gives expert-level help with a chemical, biological, radiological or nuclear weapon, carries out specified conduct without meaningful human oversight, or evades its developer's or user's control. The definition has exclusions, including publicly accessible information and lawful federal government activity.

Enforcement. A large frontier developer faces a civil penalty "that does not exceed one million dollars ($1,000,000) per violation," in an amount that depends on severity. It applies to failing to publish or transmit a required document, making a barred false statement, failing to report an incident, or not following its own framework. The penalty is "recovered in a civil action brought only by the Attorney General." The chapter holding these duties gives the public no right to sue, and the act says it does not limit rights or remedies under existing law.

Whistleblowers and local rules. Labor Code 1107.1 bars rules or retaliation that stop covered employees, meaning those responsible for assessing, managing or addressing risk of critical safety incidents, from reporting to the Attorney General, federal authorities or supervisors. A covered employee may bring a civil action under that section: the court may award reasonable attorney's fees to a successful plaintiff, and the employee may petition the superior court for temporary or preliminary injunctive relief. Large developers must also run an anonymous internal reporting process. SB 53 preempts city and county rules adopted on or after January 1, 2025, that specifically regulate frontier developers' management of catastrophic risk.

SB 942 and SB 1000: The California AI Transparency Act

The California AI Transparency Act is a content-provenance law. SB 942 (Ch. 291, signed September 19, 2024) created it, AB 853 (Ch. 674, 2025) delayed it to August 2, 2026, and SB 1000 (Ch. 861, signed September 30, 2026) rewrote it as an urgency statute with immediate effect. Many guides still describe the pre-SB 1000 version.

What SB 1000 changed.

• A "covered provider" is now a person that creates, codes or otherwise produces a generative AI system "publicly accessible within the geographic boundaries of the state." The former 1,000,000-user threshold is gone.
• The "AI detection tool" is now a "disclosure verification tool." The covered provider must offer it at no cost, and it must be publicly accessible, accept uploads or URLs, and support an interface such as an API. A provider may instead direct users to a compliant third-party tool.
• The duty to offer users a manifest disclosure option is deleted.
• The latent disclosure must now also state whether the system created or altered the content.
• The licensee rule changed. A provider that knows a licensee modified a system out of compliance has 72 hours to end the authorization or notify the licensee. The licensee then has 96 hours to report its fix, and the provider reports a failure to the Attorney General.

What the latent disclosure requires. "To the extent it is technically feasible," a covered provider must embed a disclosure in image, video or audio content its system creates or alters, other than by minor modification. It conveys the provider's name, the system's name and version, the time and date, a unique identifier, and whether the system created or altered the content.

Exclusions. The chapter does not apply to products that provide exclusively non-user-generated videogames. Before January 1, 2029, it does not apply to a system designed to primarily function as assistive technology.

Penalties. A violator owes $5,000 per violation in a civil action by the Attorney General, a city attorney or a county counsel, each day is a discrete violation, and a prevailing plaintiff recovers fees. Falsely claiming assistive-technology status carries $50,000 per violation under Section 22757.4.1, until that section is repealed January 1, 2029.

Platforms, hosting sites and devices. From January 1, 2027, large online platforms (more than 2,000,000 unique monthly users in the preceding 12 months) must detect compliant provenance data on content they distribute and give users a way to see it. AB 2713 (Ch. 856, 2026) revised that duty and clarified that platforms need not act on data that does not meet widely adopted standards. Also from that date, a GenAI hosting platform "shall not knowingly make available a GenAI system that does not place disclosures" as the act requires. From January 1, 2028, makers of capture devices first produced for sale in California on or after that date must offer a latent disclosure option and embed latent disclosures by default, to the extent technically feasible and compliant with widely adopted specifications.

Several other states, including Utah and Maine, have AI disclosure laws of their own.

AB 2013: California's AI Training Data Law

AB 2013 (Ch. 817, signed September 28, 2024) requires developers to post documentation about the data used to train generative AI systems. The duty applied "on or before January 1, 2026," and applies again before each new release or substantial modification made publicly available to Californians, for systems released on or after January 1, 2022. It applies whether or not the use is paid.

The documentation has 12 items: sources or owners, purpose, number of data points, data types, whether data is copyrighted, trademarked or patented, whether it was purchased or licensed, personal information, aggregate consumer information, cleaning or processing, collection period, first-use dates, and synthetic data. The law exempts systems used solely for security and integrity, systems whose sole purpose is operating aircraft in the national airspace, and national security, military or defense systems available only to a federal entity.

The chaptered text states no penalty and names no enforcer. xAI's challenge to the law is covered in the court cases section below.

SB 243 and Adam's Law: California Chatbot Laws

California has separate rules for companion chatbots today and a stricter child-safety regime starting in 2027. A "companion chatbot" is an AI system with a natural language interface that gives adaptive, human-like responses and can meet a user's social needs, including by sustaining a relationship across interactions. Customer service bots, certain video game bots and certain stand-alone voice assistants are excluded.

SB 243: In Force Now

SB 243 (Ch. 677, signed October 13, 2025) took effect January 1, 2026. An operator must give clear notice that the chatbot is artificial if a reasonable person would be misled into thinking it is human. An operator also may not let the chatbot engage users unless it maintains, and publishes details of, a protocol for preventing suicidal ideation, suicide or self-harm content, including crisis referrals.

For a user the operator knows is a minor, it must disclose that the user is interacting with AI, give a default reminder at least every three hours to take a break and that the chatbot is not human, and take reasonable measures against producing visual material of sexually explicit conduct or directly stating that the minor should engage in it. Operators must also disclose that companion chatbots may not be suitable for some minors.

A person "who suffers injury in fact" from a violation may sue for injunctive relief, the greater of actual damages or $1,000 per violation, and fees and costs. Annual reports to the Office of Suicide Prevention begin July 1, 2027.

SB 1119, Adam's Law: Main Duties From July 1, 2027

SB 1119 (Ch. 190, signed September 10, 2026) adds Chapter 11.6, "Adam's Law." Its risk-assessment, child-user, record-preservation, data and advertising sections become operative July 1, 2027. Section 21811 on age determination and the enforcement section carry no delayed date, so they take effect January 1, 2027, under the general rule in the California Constitution. A child is anyone under 18. An operator must determine age under state age-assurance law or apply the child protections to all users.

An operator that allows child users must, before releasing a new or substantially modified companion chatbot, perform and document a risk assessment and take measures that reasonably mitigate identified child safety risks. It must also provide:

• A crisis protocol with a referral to crisis services and, on a credible and imminent threat, parent notice or direct access to 988 or an equivalent helpline.
• Default settings only a parent can change: persistent conversational memory off, push notifications off, one hour per continuous session, and two hours per day. Stored past conversations for users 16 or older are treated differently if they meet the statute's conditions.
• Parental controls, including the ability to disable access for a child under 16.
• Repeated notices that the child is dealing with AI.
• Reasonable measures against listed conduct, such as encouraging self-harm, simulating romantic interest, claiming sentience, or encouraging reliance on the chatbot for emotional support.
• Limits on advertising, and a ban on selling a child user's personal information gathered through the chatbot.

An operator that bars child users must instead publish a high-level description of how it complies with the age-assurance section. A college offering a chatbot only for educational settings, or an entity offering one only to its own personnel for workplace settings, is not an operator under the chapter.

Child safety audits are due on or before January 1, 2029, or before first public release if later, then every two years. An operator with under $500,000,000 in prior-year gross revenue is exempt from the audit section before January 1, 2032.

A public prosecutor may seek up to $5,000 per affected child for each negligent violation and up to $15,000 for each intentional violation. A child with actual harm from a violation of specified duties, or a parent or guardian, may sue for actual damages, fees and injunctive relief. Financial harm must exceed $1,000 per child, and emotional harm must be serious emotional distress.

SB 1119 Section 1 also deletes SB 243's subdivision on users an operator knows are minors, Business and Professions Code 22602(c). Section 1 has no delayed start, so under the general rule the deletion takes effect January 1, 2027, six months before the new child-safety sections begin.

SB 867: Toys

From January 1, 2027, until January 1, 2031, no person may manufacture, sell, exchange, possess with intent to sell or exchange, or offer to a retailer a toy that includes a companion chatbot. A toy is a physical product designed, marketed or manufactured for play by children under 16. The Legislative Counsel's digest says violations carry the same civil liability as SB 243; the new section itself lists no remedy.

AI in Hiring and the Workplace: FEHA Rules, CCPA ADMT and SB 947

New York City's Local Law 144 is the only law we found in our state-by-state review that requires a bias audit of a hiring tool, as of November 4, 2026. California applies existing anti-discrimination law to automated tools by rule, and its workplace statutes begin in 2027.

Civil Rights Council ADS Regulations (In Force)

The Civil Rights Council's regulations on automated-decision systems were filed June 27, 2025, and took effect October 1, 2025. They apply the Fair Employment and Housing Act to those systems. They extend the four-year recordkeeping rule of Government Code 12946 to automated-decision system data.

They do not mandate a bias audit. Under 2 CCR 11009(f), "Relevant to any such claim or available defense is evidence, or the lack of evidence, of anti-bias testing or similar proactive efforts to avoid unlawful discrimination, including the quality, efficacy, recency, and scope of such effort, the results of such testing or other effort, and the response to the results." The Council contrasted its action with AB 2930, which would have imposed "impact assessments, pre-use notices and opt-out processes."

CCPA ADMT Regulations (Duties From January 1, 2027)

The California Privacy Protection Agency's regulations took effect January 1, 2026. A business using automated decisionmaking technology (ADMT) for a significant decision before 2027 must comply by January 1, 2027. A later user must comply any time it uses ADMT for a significant decision.

ADMT is technology that processes personal information and "uses computation to replace human decisionmaking or substantially replace human decisionmaking." A significant decision is one resulting in the provision or denial of financial or lending services, housing, education enrollment or opportunities, employment or independent contracting opportunities or compensation, or healthcare services. Insurance is not listed.

The core duties are a pre-use notice, a right to opt out, and a right to access information about the use. An opt-out is not always required: exceptions include offering an appeal to a human reviewer with authority to overturn the decision. Separate risk-assessment duties began January 1, 2026.

SB 947: Discipline and Termination (From July 1, 2027)

SB 947 (Ch. 859, signed September 30, 2026) adds Labor Code 1520 ff. It targets discipline and termination; it is not a hiring law. An employer "shall not rely solely on an ADS when making a disciplinary or termination decision." If it primarily relies on an ADS output, it must direct a human to corroborate the decision, and it may not use an output it cannot corroborate.

Employers also may not use an ADS to violate labor or civil rights law, to infer an employee's protected status, or to predict and take adverse action against a worker for exercising legal rights. A written post-use notice is required when an employer primarily relied on an ADS for such a decision, and the employee may request a description of their own data used.

The Labor Commissioner and public prosecutors enforce it, with a $500 civil penalty per violation. In those civil actions the statute allows temporary or preliminary injunctive relief, including punitive damages, and fees and costs. A business subject to the CCPA remains subject to the CPPA's ADMT regulations.

AB 1883, AB 1331 and SB 951 (From January 1, 2027)

AB 1883 (Ch. 853) bars an employer from using a workplace surveillance tool that uses AI on employees to recognize or infer an individual's emotional state, or to collect neural data. It does not stop a tool used to ensure safety, or one outside those two uses.

AB 1331 (Ch. 851) applies to workplace surveillance tools generally, not only AI. It bars monitoring in workplace bathrooms unless directed by court order, and lets employees leave tools behind when entering a bathroom, with exceptions. Both acts allow a penalty of up to $500 per violation and are enforced by the Labor Commissioner or a public prosecutor. Under AB 1883, an employee may also recover the penalty as a statutory penalty paid to the employee, though not on top of a civil penalty for the same violation. In civil actions by the Labor Commissioner or a public prosecutor under either act, the statutes also allow temporary or preliminary injunctive relief, including punitive damages, and fees and costs.

SB 951 (Ch. 860) requires a Cal/WARN notice for a layoff, relocation or termination caused in whole or in substantial part by AI or other automated technology to include the job functions being automated. The Employment Development Department will publish summaries and must report to the Legislature on AI's effects on hiring by January 1, 2028.

AI in Health Care: AB 3030, SB 1120, AB 489, AB 1979 and SB 503

AB 3030 (in force since January 1, 2025). A health facility, clinic, physician's office or group practice office that uses generative AI for patient communications about clinical information must include a disclaimer and clear instructions for reaching a human. The placement depends on the format:

The rule does not apply to a communication "read and reviewed by a human licensed or certified health care provider," and administrative matters such as scheduling and billing are outside it. Enforcement runs through the licensing articles for facilities and clinics and through the Medical Board and Osteopathic Medical Board for physicians. The chaptered text states no dollar penalty figure.

SB 1120 (in force since January 1, 2025). Health plans and disability insurers that use "an artificial intelligence, algorithm, or other software tool" for utilization review must ensure, among other things, that the tool "does not supplant health care provider decisionmaking," does not discriminate against enrollees, and is open to inspection for audit or compliance reviews.

AB 489 (in force since January 1, 2026). An AI system's advertising or functionality may not use a term implying that care or advice comes from a licensed natural person. Licensing boards enforce it, each use is a separate violation, and boards may seek an injunction or restraining order.

AB 1979 (from January 1, 2027). A business offering a "health care chatbot" to consumers for managing information, or for diagnosis or treatment, is deemed a provider of health care under the Confidentiality of Medical Information Act. Facilities, clinics and physician offices must take reasonable steps so that a licensed professional keeps independent judgment when care is informed by a clinical decision support system. They may not use AI "to independently perform any clinical function that is required by law to be performed by a person with a professional license."

SB 503 (from January 1, 2027). Business and Professions Code 22758 requires developers and deployers of clinical decision support systems to make reasonable efforts to identify systems with known or reasonably foreseeable risk of biased impacts. Developers must make reasonable efforts to mitigate that risk and give deployers documentation, and deployers must monitor and mitigate. Compliance "shall not be used as a defense to a claim of unlawful discrimination." We found no enforcement section in the chaptered text.

SB 813, AB 1405 and Executive Order N-9-26: AI Audits and the Kill Switch Proposal

California's 2026 oversight laws build the machinery for third-party AI checks. They do not by themselves require developers to use it.

SB 813 (Ch. 179, signed September 9, 2026) tells the Government Operations Agency, on or before January 1, 2028, to set application rules and criteria for "independent verification organizations," or IVOs. The act says it does not require anyone to engage an IVO or undergo a covered audit as a condition of developing or operating an AI system. In a harm lawsuit, an audit performed to a standard identified under the chapter is "relevant to, but not conclusive of, the action."

AB 1405 (Ch. 178, signed September 9, 2026) tells the agency to create an AI Auditor Registry no later than January 1, 2029. Beginning that date, a person "shall not offer, sell, or conduct a covered AI audit" unless registered. A covered AI audit is one assessing internal controls, processes or systems for an AI system or model that are necessary for compliance with state law. Registered auditors must meet independence standards, and a violation can bring removal from the registry and referral to the Attorney General or another enforcement authority.

Executive Order N-9-26 (September 18, 2026) moves the agency's dates up. It sets May 1, 2027, for the SB 813 requirements and December 1, 2027, for the AB 1405 registry steps. It also orders recommendations to the Governor by November 16, 2026, on the technical feasibility and potential efficacy of amending state law to:

• Require large frontier developers to embed IVOs onsite for periodic audits.
• Require independent verification of filed safety frameworks, transparency reports and risk assessments.
• Require a "kill switch" for frontier models, with its efficacy verified by an IVO.
• Expand "critical safety incident" to cover a range of loss-of-control incidents.

These are proposals under study, not law. The order says it does not "create any rights or benefits, substantive or procedural, enforceable at law or in equity." Earlier 2026 orders address trusted AI procurement (N-5-26, March 30) and the AI workforce (N-6-26, May 21).

AB 316 and Other California AI Laws

AB 316 (Civ. Code 1714.46, in force since January 1, 2026). In an action against a defendant who developed, modified or used AI alleged to have caused harm, "it shall not be a defense, and the defendant may not assert, that the artificial intelligence autonomously caused the harm to the plaintiff." Defendants keep other affirmative defenses, including evidence on causation or foreseeability, and evidence of others' comparative fault.

AB 325, pricing algorithms (in force since January 1, 2026). The Cartwright Act now makes it unlawful to use or distribute a common pricing algorithm as part of a contract, combination or conspiracy to restrain trade, or to coerce another person to adopt its recommended price or term.

SB 524, police reports (in force since January 1, 2026). Each law enforcement agency must maintain a policy requiring a report generated with AI to identify the program, state that it was written fully or in part using AI, and carry the officer's signature verifying the contents. Agencies keep the first draft and an audit trail for as long as the report is retained.

SB 1050, synthetic performers in ads (from January 1, 2027). It is unlawful to create and cause to be published an ad that prominently includes a synthetic performer without a clear and conspicuous disclosure. A synthetic performer is a generative AI figure or voice that gives the realistic impression of a human performer who is not recognizable as any identifiable natural person. An advertising medium is barred only after a court order is served on it.

SB 574, lawyers and arbitrators (from January 1, 2027). An attorney may not delegate the practice of law to generative AI. One who uses it must keep confidential and nonpublic information out of systems without restricted access, take reasonable steps to verify and correct its output, and disclose its use to the court for documents submitted to the court. Papers filed in court may not contain citations the responsible attorney has not personally verified. Arbitrators may not delegate decisionmaking to generative AI, and may not rely on AI-generated information outside the record without appropriate disclosure to the parties beforehand.

SB 1159, public agencies (from January 1, 2027). AI systems, autonomous agents and robots are not a "person" or "member of the public" under open-records, open-meetings and similar acts. Knowingly using AI to falsely represent that a natural person engaged with an agency is prohibited.

Digital replicas. AB 2602 (Lab. Code 927) makes certain digital replica contract terms unenforceable, and AB 1836 (Civ. Code 3344.1) covers replicas of deceased personalities. SB 1111 (Ch. 862, 2026) adds digital replicas to the voice-or-likeness statute, Civil Code 3344, and to false impersonation under the Penal Code.

State government. AB 2885 (2024) set a uniform statutory definition of AI. SB 896 (2024) and AB 979 (2025) address state government use and security, and AB 2392 (2026) creates a working group on generative AI procurement and training at public colleges.

Deepfakes and AI Note-Takers in California

California has separate statutes on election deepfakes (AB 2655, AB 2839, AB 2355) and sexually explicit deepfakes (including AB 621, which amended Civil Code 1708.86). Courts have limited two of the election laws: a federal judgment held AB 2655 preempted by Section 230 as applied to X Corp., Rumble and any such provider of an interactive computer service, with an injunction running to X Corp. and Rumble, and a federal court permanently enjoined AB 2839 against the named plaintiffs. Details are on the California deepfake laws page.

AI note-takers raise consent questions because Penal Code 632 punishes recording a confidential communication intentionally and without the consent of all parties. The California AI meeting recording laws page covers those tools.

X.AI v. Bonta and Other California AI Court Cases

X.AI LLC v. Bonta (AB 2013). xAI sued in the Central District of California (No. 2:25-cv-12295), raising Takings, First Amendment and vagueness claims. Judge Bernal denied a preliminary injunction on March 4, 2026, because the court found xAI had not established a likelihood of success on the merits. The district docket records a March 23, 2026, order on a joint stipulation to stay the case pending appeal. The Ninth Circuit's calendar, read November 4, 2026, lists argument in No. 26-1591 for November 18, 2026, in Pasadena. The appellate docket entries we read run through October 21, 2026, and show no ruling and no order blocking AB 2013.

Kohls v. Bonta (election deepfakes). On August 20, 2025, the Eastern District of California declared that AB 2655 "violates and is preempted by" 47 U.S.C. 230(c)(1), (c)(2)(B) and (e)(3), as applied to X Corp. and Rumble, and that applying it to any such provider of an interactive computer service is preempted. The permanent injunction runs to X Corp. and Rumble. In a stipulated order, the state agreed not to enforce AB 2655 against any provider of an interactive computer service, unless an appellate court vacates the judgment. The state appealed (9th Cir. No. 25-6138). The docket entries we read run through August 12, 2026, and show no decision.

Attorney General action on xAI. Attorney General Bonta announced an investigation on January 14, 2026, and sent a cease-and-desist letter on January 16, 2026, citing Civil Code 1708.86, Penal Code 311 et seq. and 647(j)(4), and the Unfair Competition Law.

What we did not find. A CourtListener search for "artificial intelligence" cases naming Attorney General Bonta, filed after September 1, 2025, showed no challenge to SB 53, SB 243, the AI Transparency Act, AB 3030 or SB 1120. That is one search on one channel, with no bill-number queries, not proof that none exists.

Vetoed and Failed California AI Bills

Bills that failed still matter, because older guides present some of them as law. None of the bills below is California law.

Two child-safety AI initiatives appear on the Secretary of State's page of measures withdrawn or failed to qualify. The Secretary of State's qualified-measures page contains no reference to artificial intelligence.

Federal Preemption and California AI Laws

As of November 4, 2026, we found no federal statute in the public laws of the 119th Congress that sets general rules for private AI systems or generally preempts state AI laws. In December 2025, Executive Order 14365 directed federal agencies to challenge or discourage state AI laws the administration views as conflicting with national policy, and called for a federal framework from Congress. The order itself does not declare any state law invalid.

California's position differs in one respect. An existing federal statute, Section 230, was the basis of the AB 2655 judgment described above; that ruling did not rest on the order. In the cases on our preemption tracker, we found no decision holding a state AI law preempted by an executive order.

What Changes in California AI Law in 2027

California's 2026 session produced most of the dates below. Acts with no urgency or delayed-start clause in their chaptered text take effect January 1, 2027, under the general rule in the California Constitution, art. IV, sec. 8(c). The other dates come from each law's own text, and CPPA section numbers follow the approved regulation text.

Pending bills. Floor action in the 2025-26 regular session ended August 31, 2026, and the session adjourns November 30, 2026. We found no bill for the next session to list as of November 4, 2026. The docket entries we read show no decision in Ninth Circuit No. 26-1591 (through October 21, 2026) or No. 25-6138 (through August 12, 2026).

More California Laws

• California Recording Laws
• California Data Privacy Laws
• California Surveillance Camera Laws
• California Background Check Laws
• California Whistleblower Laws
• California Deepfake Laws
• California AI Meeting Recording Laws
• State and Federal AI Laws Guide

Frequently Asked Questions

Does California have an AI law?

Yes, several, though no single comprehensive AI act as of November 4, 2026. California regulates AI by subject, through laws including SB 53 (frontier models), the California AI Transparency Act (SB 942, as amended), SB 243 (companion chatbots), AB 2013 (training data) and health care and employment rules.

What is California SB 53?

SB 53 is the Transparency in Frontier Artificial Intelligence Act, in force since January 1, 2026. It requires developers of frontier models trained with more than 10^26 operations to publish transparency reports and report critical safety incidents. Developers with more than $500 million in annual revenue must also publish a frontier AI framework. Only the Attorney General can seek the civil penalty, of up to $1,000,000 per violation, against a large frontier developer. Covered employees have a separate whistleblower action.

Is the California AI Transparency Act in effect?

Yes. It became operative August 2, 2026, and SB 1000 rewrote it with immediate effect on September 30, 2026. Large online platform and hosting platform duties start January 1, 2027, and capture device duties start January 1, 2028.

Does SB 942 still have a 1,000,000-user threshold?

No. SB 1000 removed the threshold on September 30, 2026. A covered provider is now a person that creates, codes or otherwise produces a generative AI system publicly accessible in California.

What is California's AI chatbot law?

SB 243, in force since January 1, 2026, requires companion chatbot operators to give notice where a reasonable person would be misled and to maintain and publish a self-harm protocol. For users known to be minors it adds AI disclosure, three-hour break reminders and measures against sexually explicit material; SB 1119 deletes that subdivision effective January 1, 2027, under the general rule. SB 1119, Adam's Law, adds child-safety duties, mostly from July 1, 2027.

Can you sue a chatbot company under California law?

SB 243 lets a person who suffers injury in fact from a violation sue for injunctive relief, the greater of actual damages or $1,000 per violation, and fees. From July 1, 2027, SB 1119 adds a narrower action for a child with actual harm from specified violations.

What is Adam's Law in California?

Adam's Law is SB 1119, signed September 10, 2026. From July 1, 2027, companion chatbot operators that allow child users must assess child safety risks and apply crisis protocols, parent-only default limits and advertising limits. Operators that bar child users must publish how they comply with the age-assurance section. Independent child safety audits are due by January 1, 2029, with an exemption for smaller operators before 2032.

Does California require bias audits for AI hiring tools?

No statute or rule we found requires one. Under the Civil Rights Council's regulations, in force since October 1, 2025, anti-bias testing or its absence is relevant evidence in a FEHA claim or defense. In our state-by-state review, the only such law we found is New York City's Local Law 144.

When do California's ADMT rules take effect?

The CCPA regulations took effect January 1, 2026, and the ADMT duties apply from January 1, 2027, to businesses using ADMT for significant decisions on lending, housing, education, employment or health care.

Does California require doctors to disclose AI use?

AB 3030 requires health facilities, clinics and physician offices to disclose generative AI use in patient communications about clinical information and to explain how to reach a human. It does not apply to a communication read and reviewed by a licensed or certified provider.

Did Governor Newsom veto an AI bill?

Yes. He vetoed SB 1047 on September 29, 2024, SB 7 and AB 1064 on October 13, 2025, and AB 2575, AB 2656 and SB 903 in September 2026, among others.

Is California's AI training data law blocked?

We found no order blocking it in the docket entries we read, which run through October 21, 2026. A federal judge denied xAI's request for a preliminary injunction against AB 2013 on March 4, 2026, and Ninth Circuit argument is set for November 18, 2026.

Does Trump's AI executive order cancel California's AI laws?

Executive Order 14365 does not itself declare any state law invalid. It directs federal agencies to challenge or discourage state AI laws the administration views as conflicting with national policy. We found no decision on our preemption tracker holding a state AI law preempted by an executive order as of November 4, 2026.