Artificial intelligence technology.

The bill text was reformatted from a formal enrolled document (with Senate/Assembly passage dates, signature lines for Secretary of the Senate, Chief Clerk of the Assembly, and Governor's receipt) into a web page layout with navigation elements, version history dropdown, and publication metadata. No changes to any substantive provision.

The civil penalty for failing to display the required consumer warning was reduced from $25,000 per day to $10,000 per day.

A new exemption was added: the consumer warning is not required for a digital replica created in a video game where the replica is used solely in gameplay and is not distributed outside of the game. A definition of 'video game' referencing Civil Code Section 1746 was also added.

A new Section 2.5 was added that incorporates SB 683's amendments to Civil Code Section 3344, which include: (1) a right to seek an injunction or temporary restraining order with a two-business-day deadline for the respondent to remove or recall content, and (2) a rebuttable presumption that using an employee's photograph incidentally in an advertisement is not a knowing unauthorized use. This section only becomes operative if both SB 11 and SB 683 are enacted by January 1, 2026, and SB 11 is enacted after SB 683.

The consumer warning must now also be included in the terms and conditions for use of the technology, in addition to being displayed before first use and hyperlinked on prompt input screens.

A new Section 5 was added specifying the conditions under which Section 2.5 becomes operative: both bills must be enacted and effective on or before January 1, 2026, each must amend Section 3344, and SB 11 must be enacted after SB 683 (in which case Section 2 of SB 11 does not become operative).

Changed the covered activity from 'sells or provides access to' AI technology to 'makes available to consumers,' which is a broader standard that could capture more providers and distribution models.

Changed the technology threshold from being 'capable of creating any digital replica' to 'enables a user to create a digital replica,' broadening coverage to include tools that facilitate user creation rather than only those directly generating replicas.

Removed the requirement for the Department of Consumer Affairs to determine the acceptable form and content of the consumer warning and post it on its website by July 1, 2026. The warning text is now specified directly in the statute.

Changed enforcement authority from the Department of Consumer Affairs to 'a public prosecutor' who may bring a civil action in any court of competent jurisdiction.

Removed the provision that civil penalties collected under the warning requirement shall be deposited into the General Fund.

Changed the warning language from 'misuse of the technology may result in civil or criminal liability' to 'unlawful use of the technology to depict another person without prior consent may result in civil or criminal liability,' making the warning more specific about what conduct is prohibited.

The description of covered AI technology was simplified from 'that is capable of creating any enables a user to create a digital replica' to 'that enables a user to create a digital replica,' removing redundant and awkward language while keeping the same substantive meaning.

A new definition for 'public prosecutor' was added, specifying it means the Attorney General, a city attorney, county counsel, district attorney, or any other city or county prosecutor. This clarifies who has authority to enforce the consumer warning requirement under Section 1(c).

The definitions in Section 1 were restructured from separate subsections (d) and (e) into a single consolidated subsection (d) with numbered items (1) through (3), improving organization without changing the defined terms.

The definition of 'synthetic content' (covering text, audio, videos, images, and deepfakes generated or significantly modified by algorithms) is entirely removed from the bill. All references to synthetic content are replaced with 'digital replica,' which is limited to computer-generated representations of a specific individual's voice or visual likeness.

The consumer warning requirement in Section 22650 now applies only to AI technology designed to create a 'digital replica' rather than any 'synthetic content.' This significantly narrows which AI products must carry the warning - for example, an AI text generator or image generator that does not replicate a specific person's likeness would no longer be covered.

Definitions of 'artificial intelligence' and 'digital replica' are restructured to cross-reference Civil Code Section 3110 (for AI) and Section 3344.1 (for digital replica) rather than being defined solely within the Penal Code. The AI definition is also slightly reworded, removing 'may operate with varying levels of autonomy' and restructuring the sentence.

A new subdivision (f) is added to Civil Code Section 3344 stating that 'a voice or likeness includes a digital replica, as defined in Section 3344.1,' making explicit that the civil right of publicity statute covers AI-generated replicas.

The new chapter in the Business and Professions Code is renumbered from Chapter 22.7 to Chapter 22.6, likely to align with existing code structure or avoid a numbering conflict.

The covered-entity language changed from 'sells or provides access to any artificial intelligence technology that is designed to create' to 'sells or provides access to makes available to consumers any artificial intelligence technology that is capable of creating.' This broadens coverage in two ways: it adds 'makes available to consumers' as a trigger, and replaces 'designed to create' with the broader 'capable of creating,' so technology does not need to be specifically designed for digital replica creation to be covered.

A new legislative finding was added referencing the Unfair Competition Law's statutory cause of action and civil penalties, providing additional legal justification for the bill's penalty provisions.

The consumer warning text was changed from a general statement about 'misuse of the technology' to a specific statement: 'Unlawful use of this technology to depict another person without prior consent may result in civil or criminal liability for the user.' This narrows and clarifies what the warning must communicate.

A new subdivision (b) requires that the warning be displayed to the consumer before their first use of the technology and thereafter be hyperlinked from any page or screen where the consumer may input a prompt. All warnings must be displayed in a manner that is clear and conspicuous.

Enforcement authority was shifted from 'the department' (Department of Consumer Affairs) to 'a public prosecutor,' meaning district attorneys or city attorneys can now bring civil actions for violations rather than only the state agency.

A new definition of 'digital replica' is added to Penal Code Section 540(b)(1): a computer-generated, highly realistic electronic representation readily identifiable as the voice or visual likeness of an individual in a recording, image, or transmission where the person did not actually perform/appear or the fundamental character was materially altered. An exclusion is added for authorized reproductions, samples, remixing, mastering, or digital remastering by copyrightholders.

The definition of 'artificial intelligence' in Penal Code Section 540(a) is restructured. The introduced version described AI as a system that 'may operate with varying levels of autonomy' (permissive language). The amended version describes it as a system that 'varies in its level of autonomy' (descriptive), making autonomy variation a defining characteristic rather than an optional feature.

Throughout the bill, references to 'synthetic content' are expanded to also cover 'digital replica.' This affects the consumer warning requirement (Business and Professions Code 22650), the civil rights provision (Civil Code 3344(f)), and the criminal false personation provision (Penal Code 541), broadening which AI-generated content triggers legal protections.

The rebuttable presumption in Civil Code Section 3344 that failure to obtain an employee's consent was not a 'knowing use' when the employee's photo appeared incidentally in a publication is removed. This eliminates a defense for employers and publishers who used employee images without explicit consent.

The Business and Professions Code chapter is renumbered from Chapter 22.7 to Chapter 22.6, likely to accommodate another bill or correct a codification conflict.

All definitions of 'artificial intelligence' and 'digital replica' across the bill (Penal Code Section 540, Business and Professions Code Section 22650, Evidence Code Section 1425) were changed from standalone definitions or cross-references to Penal Code Section 540 to cross-references to Civil Code Sections 3110 and 3344.1. This consolidates the definitional authority in one place.

Civil Code Section 3344(f) was changed from requiring that a digital replica be one 'a reasonable person would believe to be the genuine voice or photograph' to simply stating that 'a voice or likeness includes a digital replica.' This removes the reasonable-person threshold test and broadens coverage from 'voice or photograph' to 'voice or likeness,' making it easier for a plaintiff to bring a claim based on an AI-generated replica.

Penal Code Section 541 was reworded from 'the use of a digital replica with the intent to impersonate another is deemed to be a false personation' to 'false impersonation includes the use of a digital replica with the intent to impersonate another is deemed to be a false personation.' The new language frames digital replica use as an inherent part of the false impersonation concept rather than a separate statutory deeming provision.

The legislative counsel's digest was updated to reflect the reworded criminal and civil provisions, and page numbering was adjusted throughout the bill.

Penal Code Section 541 was simplified. The Senate version stated that false impersonation includes use of a digital replica with intent to impersonate, that this is 'deemed to be a false personation,' and that intent can be inferred if the replica would lead a reasonable person to believe it is genuine. The Assembly version removed both the 'deemed to be a false personation' language and the reasonable person inference standard, leaving only the basic statement that false impersonation includes use of a digital replica with intent to impersonate.

The Judicial Council's mandate under Section 1425 was expanded. The Senate version required the Council to review AI's impact on introduction of evidence and develop rules to help courts assess whether evidence was generated or manipulated by AI. The Assembly version added 'admissibility' and 'proffered' language and explicitly requires the Council to also determine whether such evidence is admissible, giving the mandate a broader evidentiary focus.

The consumer warning requirement in Section 22650 was broadened. The Senate version applied to AI technology 'designed to create' a digital replica. The Assembly version changed this to 'capable of creating,' which would cover a wider range of AI tools that can produce digital replicas even if not specifically designed for that purpose. Note: the final text contains both phrases ('designed to create capable of creating'), suggesting a drafting error.

The Assembly version clarified 'The department' to 'The Department of Consumer Affairs' in Section 22650(b), removing ambiguity about which agency is responsible for specifying the consumer warning form and content.