WHAT THIS BILL REGULATES · 1 REQUIREMENT TYPE
How Is This Bill Enforced
Verbatim statutory text on the left; plain-language analysis and a per-section checklist on the right. Numbered markers cross-link to the matching checklist row.
This Act may be cited as the Provenance DataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5 Requirements Act.
Designates the short title of the Act as the Provenance Data Requirements Act. Creates no compliance obligation.
As used in this Act: [definitions of Artificial intelligenceArtificial intelligence"Artificial intelligence" means a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs, such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. For the purposes of this definition a system that uses "artificial intelligence" may vary in its level of autonomy and adaptations after deployment.Section 5, Capture deviceCapture device"Capture device" means a device that can record photographs, audio, or video content, including, but not limited to, video and still photography cameras, mobile phones with built-in cameras or microphones, and voice recorders.Section 5, ContentContent"Content" means any visual or audio digital content.Section 5, Generative artificial intelligence modelGenerative artificial intelligence model"Generative artificial intelligence model" means an artificial intelligence model designed to generate new data or content based on the patterns or structures from its training data. "Generative artificial intelligence model" does not include nongenerative capabilities like classification, predictions, or labeling.Section 5, Generative artificial intelligence tool providerGenerative artificial intelligence tool provider"Generative artificial intelligence tool provider" means an entity that develops a generative artificial intelligence tool that is made publicly available for use by people in this State.Section 5, Generative artificial intelligence toolGenerative artificial intelligence tool"Generative artificial intelligence tool" means a product or feature that provides the outputs from a generative artificial intelligence model to end users.Section 5, Large online platformLarge online platform"Large online platform" means a public-facing social media platform, content-sharing platform, messaging platform, advertising network, or standalone search engine that distributes content to users who did not create or collaborate in creating the content. For the purposes of this definition, a "large online platform" shall have 1,000,000 or more unique monthly users during the preceding 12-month period.Section 5, Provenance application toolProvenance application tool"Provenance application tool" means a tool or online service that makes provenance data difficult to remove and allows users to easily apply provenance data.Section 5, Provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5, Provenance readerProvenance reader"Provenance reader" means a tool or service that allows users to view provenance data and its integrity.Section 5, and Synthetic contentSynthetic content"Synthetic content" means visual or audio content that has been produced or modified by a generative artificial tool.Section 5 as set forth verbatim in the top-level definitions dictionary].
Section 5 defines the operative terms of the Act, including the three covered actor or product classes — generative artificial intelligence tool provider, large online platform, and capture device — and the technical anchor concept of provenance data, which is tied to widely-adopted standards-body specifications. The standards-body framing accommodates C2PA Content Credentials and similar interoperable schemes without naming any specific standard.
The definitions section creates no independent compliance obligation; each defined term is operationalized in Sections 10, 15, and 20.
(a) 1 A generative artificial intelligence tool providerGenerative artificial intelligence tool provider"Generative artificial intelligence tool provider" means an entity that develops a generative artificial intelligence tool that is made publicly available for use by people in this State.Section 5 shall apply provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5, either directly or through the use of third-party technology, to wholly-generated synthetic contentSynthetic content"Synthetic content" means visual or audio content that has been produced or modified by a generative artificial tool.Section 5 generated by the provider's generative artificial intelligence toolGenerative artificial intelligence tool"Generative artificial intelligence tool" means a product or feature that provides the outputs from a generative artificial intelligence model to end users.Section 5.
(b) 2 A generative artificial intelligence tool providerGenerative artificial intelligence tool provider"Generative artificial intelligence tool provider" means an entity that develops a generative artificial intelligence tool that is made publicly available for use by people in this State.Section 5 shall make available a provenance application toolProvenance application tool"Provenance application tool" means a tool or online service that makes provenance data difficult to remove and allows users to easily apply provenance data.Section 5 that enables the user to apply provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5, either directly or through the use of third-party technology, to contentContent"Content" means any visual or audio digital content.Section 5 that has been significantly modified to include synthetic contentSynthetic content"Synthetic content" means visual or audio content that has been produced or modified by a generative artificial tool.Section 5.
(c) 3 A generative artificial intelligence tool providerGenerative artificial intelligence tool provider"Generative artificial intelligence tool provider" means an entity that develops a generative artificial intelligence tool that is made publicly available for use by people in this State.Section 5 shall make a provenance readerProvenance reader"Provenance reader" means a tool or service that allows users to view provenance data and its integrity.Section 5 available to the public.
Section 10 imposes three distinct provenance obligations on generative AI tool providers: (a) automatic application of provenance data to wholly-generated synthetic content; (b) availability of a provenance application tool enabling users to apply provenance data to content significantly modified to include synthetic content; and (c) public availability of a provenance reader so anyone can view and verify provenance data integrity.
The (a) obligation is automatic and runs to the developer's own outputs; (b) is a tooling availability requirement covering user-driven hybrid content; (c) is a public-tooling obligation that effectively forces interoperable verification capability. The standards-compliance framing in the definition of provenance data functions as an interoperability requirement equivalent to T-02.3.
(1)–(2) 4 A large online platformLarge online platform"Large online platform" means a public-facing social media platform, content-sharing platform, messaging platform, advertising network, or standalone search engine that distributes content to users who did not create or collaborate in creating the content. For the purposes of this definition, a "large online platform" shall have 1,000,000 or more unique monthly users during the preceding 12-month period.Section 5 shall: (1) retain any available provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5 in contentContent"Content" means any visual or audio digital content.Section 5 provided to or posted on the large online platformLarge online platform"Large online platform" means a public-facing social media platform, content-sharing platform, messaging platform, advertising network, or standalone search engine that distributes content to users who did not create or collaborate in creating the content. For the purposes of this definition, a "large online platform" shall have 1,000,000 or more unique monthly users during the preceding 12-month period.Section 5; and (2) make available to a consumer of contentContent"Content" means any visual or audio digital content.Section 5 in this State either (i) the provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5 or a conspicuous indicator that provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5 is available.
Section 15 imposes two duties on large online platforms (≥1M monthly users): (1) preserve any available provenance data in content provided to or posted on the platform, and (2) surface to consumers either the provenance data itself or a conspicuous indicator that provenance data is available. The retention duty maps closely to T-02.6 (do not strip provenance signals), and the user-disclosure duty maps to T-02.5 (user-facing interface for provenance).
5 A manufacturer of a capture deviceCapture device"Capture device" means a device that can record photographs, audio, or video content, including, but not limited to, video and still photography cameras, mobile phones with built-in cameras or microphones, and voice recorders.Section 5 shall include the ability for a user to enable the inclusion of provenance dataProvenance data"Provenance data" means information, such as the origin of a piece of content and the history of modifications to the content, that is in a format that is compliant with widely-adopted guidelines or specifications promulgated by an established standards-setting body, and includes, but is not limited to: (1) whether some or all of the content is synthetic content; and (2) when there is synthetic content, the name of the generative artificial provider whose generative artificial tool generated the synthetic content.Section 5 in the user's captured contentContent"Content" means any visual or audio digital content.Section 5.
Section 20 imposes a single hardware-side obligation on capture device manufacturers: include functionality enabling a user to turn on provenance data in captured content. The duty is opt-in ("enable the inclusion") rather than default-on, distinguishing it from peer bills like California AB 853 that require latent provenance disclosures by default.