WeVote
Log In

Bill

Bill

AB 2713

California AI Transparency Act: system provenance data.

2025-2026 Regular Session Introduced by Cecilia Aguiar-Curry and 1 co-sponsor

AB 2713 strengthens California’s AI transparency by boosting penalties and requiring large platforms to let users inspect system provenance data for AI-generated or modified conten

Read second time. Ordered to third reading.
#californiaaitransparencyact #systemprovenancedata
0
WeVote Research Nonpartisan
Bill Summary · AB 2713

Summary of AB 2713 (California AI Transparency Act: system provenance data: inspection)

Overview

  • Jurisdiction: California
  • Session: 2025–2026
  • Prime sponsor: Assembly Member Buffy Wicks (co-sponsor)
  • Bill number: AB 2713
  • Purpose: Amend the California AI Transparency Act to strengthen civil penalties and to expand access to system provenance data (AI provenance data) for content produced by or distributed with generative AI (GenAI) or capture devices on large online platforms.

1) Main purpose and intent

  • Increase the accountability framework for entities that operate large online platforms and GenAI-related content in California.
  • Ensure transparency by requiring accessible inspection of system provenance data attached to content, so users can assess authenticity, origin, and history of modification.
  • Elevate civil penalties for violations to deter noncompliance.

2) Key provisions and changes

Civil penalties

  • The bill increases the civil penalty for violations of the California AI Transparency Act from $5,000 per violation to $5,500 per violation.
  • Penalties are enforceable via civil actions brought by the Attorney General, a city attorney, or a county counsel.
  • Each day a violation persists counts as a separate, discrete violation.
  • Prevailing plaintiffs may recover reasonable attorney’s costs and fees.

System provenance data inspection (large online platforms)

  • Large online platforms must enable users to inspect all available system provenance data that complies with widely adopted specifications (as set by an established standards-setting body).
  • Inspection methods (each considered acceptable):
    1. Direct inspection of system provenance data through the platform’s user interface.
    2. Allow users to download a version of the content with its attached system provenance data (subject to federal copyright law).
    3. Provide a link to the content’s system provenance data displayed on a website or in another application (provided by the platform or a third party).
  • The user interface must clearly and conspicuously disclose:
    • Whether provenance data is available.
    • The name of the GenAI system or capture device that created or substantially altered the content (if applicable).
    • Whether any digital signatures are available.
  • The platform should, to the extent technically feasible, not knowingly strip system provenance data or digital signatures from content.
  • The inspection requirement takes effect on January 1, 2027.

3) Who or what is affected

Entities subject to the act

  • Large online platforms that host or distribute content and that are capable of embedding or attaching system provenance data to content (as defined by current California AI Transparency Act standards).
  • Content creators using GenAI systems or capture devices that generate or substantially alter content distributed on large online platforms.
  • Third-party licensees (as applicable under the Act) who may be involved in violations and who could face civil actions by the Attorney General, a county counsel, or a city attorney.

Individuals and users

  • California users and the public benefit from enhanced transparency about the authenticity and provenance of AI-generated or modified content.

4) Procedural and timeline aspects

  • Effective date for most provisions (except general operative dates): January 1, 2027 for the provenance data inspection requirements.
  • The bill modifies operative sections of the Business and Professions Code (Sections 22757.3.1 and 22757.4).
  • The California AI Transparency Act currently envisages an operative date of August 2, 2026; AB 2713 adds a separate operational timeline for provenance data inspection (2027) and adjusts penalties accordingly.
  • Legislative history indicates committee amendments and standard bill processing, with recent actions in 2026 (amendments, re-referrals, and committee votes).

5) Summary of potential impact

  • Strengthened deterrence: Higher civil penalties (from $5,000 to $5,500 per violation) increase potential costs of noncompliance.
  • Enhanced transparency: Users gain practical access to system provenance data through multiple mechanisms, improving verification of authenticity and origin of AI-generated or modified content.
  • Compliance burden: Large platforms will need to implement user interfaces and data-access options for provenance data and ensure compatibility with widely adopted provenance standards.
  • Copyright considerations: Providing downloadable provenance data or links must respect federal copyright laws.

Overall, AB 2713 aims to reinforce California’s AI transparency regime by tightening penalties and making provenance data more readily inspectable by users, with a staged implementation timeline.

Compiled from official sources — confirm details with the bill’s official record.

Sign in to ask a question.