California has passed more AI-specific legislation than any other state, and most of the major rules took effect on January 1, 2026. California’s AI laws now govern when a company has to tell you that you’re interacting with a machine, how AI-generated images and voices must be labeled, what rights you have over automated decisions affecting your job or credit, how performers’ digital replicas can be used, and what safety duties apply to the largest frontier model developers. The common thread is transparency: knowing when AI is involved, and in what.
When Someone Has to Tell You It’s AI
Several California statutes exist for a single reason: so a person on the other end of a message, call, or conversation knows a machine is talking.
Bots Online
Senate Bill 1001 has been in force since July 2019. It is unlawful to use a bot to communicate with a person in California online with intent to mislead them about the bot’s artificial identity, if the purpose is to push a commercial transaction or influence a vote. Disclosing that it’s a bot cures the violation. The disclosure has to be clear, conspicuous, and reasonably designed to inform the person they aren’t talking to a human.1California Legislative Information. California Business and Professions Code 17940-17943 Bots
AI Voices on Phone Calls
Assembly Bill 2905 addresses telemarketing calls that use AI-generated voices. When an automatic dialing-announcing device is used, an unrecorded natural voice must first inform the person being called if the prerecorded message uses an artificial voice. The statute defines “artificial voice” as one generated or significantly altered using AI, and violations of the implementing Public Utilities Commission rules are treated as criminal offenses.2California Legislative Information. California Code AB-2905 Telecommunications Automatic Dialing-Announcing Devices Artificial Voices
Companion Chatbots
SB 243, effective January 1, 2026, regulates chatbots designed for ongoing interaction and emotional support. Operators must clearly notify users that they’re interacting with AI. For users the operator knows are minors, reminders have to appear at least every three hours during continuing interactions. Operators must also maintain protocols for preventing the chatbot from producing content encouraging self-harm, and when a user expresses suicidal ideation, the chatbot must refer them to a crisis service. Anyone injured by a violation can sue for injunctive relief and damages of at least $1,000 per violation, plus attorney’s fees.
AI Impersonating Healthcare Providers
Also effective January 1, 2026, Assembly Bill 489 prohibits developers and deployers of AI systems from using terms, letters, or design elements that imply the AI holds a healthcare license or that a licensed human professional is providing the care when one is not. Each use of a prohibited term counts as a separate violation. Enforcement goes to the relevant healthcare licensing board, which can seek injunctions.3California State Assembly. AB 489 Bonta Assembly Committee on Privacy and Consumer Protection Analysis
Deepfakes in Elections
California took two routes against election-related AI deception. Assembly Bill 2839 bars knowingly distributing materially deceptive audio, visual, or video content within 120 days before an election and up to 60 days after. The content has to be spread with malice and intended either to harm a candidate’s reputation or deceive voters about the electoral process. Targeted candidates can seek injunctions, civil penalties, and attorney’s fees.4California Legislative Information. California Code AB-2839 Elections Deceptive Media in Advertisements
Most of AB 2839 is not currently being enforced. A federal judge issued a preliminary injunction, ruling it a content-based speech restriction that fails strict scrutiny under the First Amendment. The court said California’s interest in election integrity “does not give legislators unbridled license to bulldoze over the longstanding tradition of critique, parody, and satire protected by the First Amendment.” Only an audio-only disclosure requirement for recordings survived.
Assembly Bill 2655, the Defending Democracy from Deepfake Deception Act, puts the obligation on large online platforms instead of individual speakers. During the 120 days before Election Day, platforms must block materially deceptive election-related content. For content depicting election officials, that duty extends 60 days after the election. Platforms must label certain digitally altered content as inauthentic and provide a reporting mechanism for California residents. The Attorney General, district attorneys, and city attorneys can bring enforcement actions, and private individuals can seek injunctive relief if a platform doesn’t respond.5LegiScan. California AB2655 2023-2024 Regular Session Amended
Your Face and Voice as Digital Replicas
Two laws address AI-generated replicas of real people in entertainment work.
Assembly Bill 2602, now Labor Code section 927, applies to living performers. A contract provision allowing a company to use a digital replica in place of a performer’s live work is unenforceable unless the provision includes a reasonably specific description of the intended uses and the performer was represented either by an attorney who negotiated the digital replica terms or by a union whose collective bargaining agreement addresses replica use. Vague “we may use your digital likeness” language cannot be enforced without those protections.6California Legislative Information. California Code Labor Code 927
Assembly Bill 1836 extends protection to deceased performers’ estates. Under revised Civil Code section 3344.1, producing, distributing, or making available a digital replica of a deceased person’s voice or likeness in an audiovisual work or sound recording without prior consent from the estate creates liability for the greater of $10,000 or actual damages. These rights last 70 years after death. Exemptions cover news, public affairs, sports broadcasts, commentary, criticism, scholarship, satire, parody, documentaries and biographical works, and fleeting or incidental uses.7California Legislative Information. California Code Civil Code 3344.1 Deceased Personalitys Likeness
Labeling AI-Generated Content
Senate Bill 942, the California AI Transparency Act, requires AI developers to build provenance information into the content their systems generate. Every AI-generated image, video, or audio file must carry a latent disclosure, meaning metadata identifying it as AI-created, to the extent technically feasible and reasonable. Developers must also offer users a manifest disclosure option: a visible, clear, conspicuous label identifying content as AI-generated. Covered providers have to make a free, publicly accessible AI detection tool available so anyone can check whether a piece of media came from their system.8California Legislative Information. California Code SB-942 California AI Transparency Act
Violations carry a $5,000 civil penalty per day of noncompliance, and each day counts as a separate violation. The Attorney General, a city attorney, or a county counsel can bring the action, and prevailing plaintiffs recover attorney’s fees and costs.9LegiScan. California Business and Professions Code 22757 California AI Transparency Act
What Went Into the Model
Assembly Bill 2013 requires generative AI developers to disclose their training data. Beginning January 1, 2026, any developer of a generative AI system available to Californians must post a high-level summary on its website covering the sources or owners of the datasets, the types of data included, the approximate number of data points, how the data supports the system’s intended purpose, and the collection time period. Developers must also disclose whether their datasets contain personal information or aggregate consumer information as defined under the California Consumer Privacy Act.10California Legislative Information. California Code AB-2013 Generative Artificial Intelligence Training Data Transparency
The requirement applies retroactively to any generative AI system released on or after January 1, 2022, and documentation must be updated before each substantially modified system is made publicly available. “Training” is defined broadly to include testing, validation, and fine-tuning.
Automated Decisions About You
On July 24, 2025, the California Privacy Protection Agency adopted regulations implementing consumer rights over automated decision-making technology, or ADMT. The rules took effect January 1, 2026. They give consumers the right to access and opt out of businesses’ use of automated systems that process personal information to make or facilitate decisions.11California Privacy Protection Agency. CCPA Updates Cybersecurity Audits Risk Assessments Automated Decisionmaking Technology and Insurance Regulations
Covered decisions include those affecting financial services, housing, insurance, education, employment, and healthcare. The opt-out right also reaches profiling of employees and job applicants using tools like keystroke loggers, productivity monitors, facial recognition, and location trackers, and profiling of consumers in publicly accessible places through Bluetooth tracking, geofencing, or license-plate recognition. AI-driven hiring tools, lending algorithms, and workplace surveillance systems fall under the CCPA’s consumer protection framework as a result.
Safety Rules for the Largest AI Developers
Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act, applies to “frontier models,” defined as foundation models trained using more than 10²⁶ integer or floating-point operations. That threshold reaches only the most powerful systems from a handful of companies.12California Legislative Information. California Code SB-53 Transparency in Frontier Artificial Intelligence Act
SB 53 separates all frontier developers from “large” frontier developers with annual gross revenues above $500 million. Large developers must publish an annual framework explaining how they identify, assess, and mitigate catastrophic risks, aligned with recognized standards such as the NIST AI Risk Management Framework. Before deploying a new or substantially modified frontier model, they also have to publish a transparency report summarizing their risk assessments and any third-party evaluation.
“Catastrophic risk” is defined as a foreseeable risk that a frontier model could contribute to the death or serious injury of more than 50 people, or cause over $1 billion in property damage, through scenarios like helping create a chemical, biological, radiological, or nuclear weapon; autonomously committing cyberattacks or violent crimes without meaningful human oversight; or evading its developers’ or users’ control.13LegiScan. California SB53 2025-2026 Regular Session Enrolled
Frontier developers must report critical safety incidents to the California Office of Emergency Services within 15 days of discovery, or within 24 hours if the incident poses an imminent risk of death or serious physical injury. Developers must also maintain anonymous internal whistleblower channels for employees to report catastrophic risk concerns and cannot retaliate against workers who raise those concerns to the Attorney General or other authorities. Civil penalties reach up to $1 million per violation.
Who Enforces These Laws and What It Costs
The California Attorney General is the primary enforcer across most of these statutes and has already sent cease-and-desist demands to AI companies over content violating state law.14State of California Department of Justice. Attorney General Bonta Sends Cease and Desist Letter to xAI Demands It Halt Illegal Actions Immediately
Penalty structures vary. SB 53 allows up to $1 million per violation for frontier safety failures. SB 942 imposes $5,000 for each day of noncompliance with transparency labeling. AB 2655 and AB 2839 allow courts to award injunctive relief and attorney’s fees to prevailing plaintiffs. Several laws, including SB 243 for companion chatbots and the deepfake statutes, create private rights of action, so affected individuals can sue directly rather than waiting for a government agency.