Personal rights: liability: social media platforms.

The bill was reformatted from a traditional enrolled legislative document into a web publication layout. This added website navigation elements (search, FAQ, sitemap), a version history list showing all amendment dates, and removed the formal chapter heading, page numbers, and governor's signature block. The actual statutory language, penalty amounts, definitions, and operative date remain unchanged.

Added an operative date of January 1, 2027, both in the bill summary and as new Section 3273.74 within the Civil Code title. This delays when the civil penalty provisions take effect.

Section 3273.73(a) was reworded from 'violates Section 51.7, 51.9, 52, or 52.1, including 52.1 through its algorithms that relay content to users' to 'violates Section 51.7, 51.9, 52, or 52.1 through its algorithms that relay content to users.' This removes the specific callout of Section 52.1 and the word 'including,' making the algorithmic liability language apply more broadly across all listed sections rather than singling out 52.1.

A new legislative finding (subsection b(3)) was added citing Department of Justice data showing anti-Jewish bias events rose 52.9 percent and anti-Islamic bias events rose 62 percent in 2023, broadening the stated justification beyond the previously cited anti-immigrant, anti-LGBTQ+, antisemitic, and anti-women findings.

The bill summary description was reworded from 'including rights through its algorithms that relay content to users' to 'through its algorithms that relay content to users,' removing the phrase 'including rights.'

Four additional Assembly members (Harabedian, Kalra, Stefani, and Zbur) were added as coauthors.

The operative liability provision (Section 3273.73) no longer references Penal Code Section 31 (terrorism) or Section 422.6 (hate crimes). Platforms are now only liable for violations of Civil Code Sections 51.7, 51.9, 52, or 52.1, significantly narrowing the legal basis for penalties.

A new statement was added to the legislative findings clarifying that 'the purpose of this act is not to regulate speech or viewpoint but to clarify that social media platforms, like all other businesses, may not knowingly use their systems to promote, facilitate, or contribute to conduct that violates state civil rights laws.' This appears designed to address First Amendment concerns.

References to 'artificial intelligence or algorithm' were reduced to just 'algorithm' in both the provision allowing algorithms to be treated as independent acts of the platform and the provision deeming platforms to have actual knowledge of their systems. This narrows the specific AI-related language while keeping the broader 'algorithm' term.

The phrase 'in an action brought pursuant to this section' was removed from the civil penalty liability language, making the penalty provision slightly less procedural in its framing.

The legislative findings were restructured: case law references (Sindell v. Abbott Laboratories, Summers v. Tice, Prosser and Keeton on Torts, Restatement Second of Torts) were removed from the findings section, and the statistical evidence was reorganized into numbered subsections with a new introductory framing about California's 'compelling interest.'

Four additional Assembly members (Harabedian, Kalra, Stefani, and Zbur) were added as coauthors of the bill.

The entire amendment to Education Code Section 48901.8 was removed. This provision would have required school districts, county offices of education, and charter schools to adopt policies limiting or prohibiting pupil use of social media at schoolsites or under school supervision. Its removal eliminates the state-mandated local program on schools.

The bill's stated purpose in the digest was narrowed. It no longer references 'Pupils: use of social media' or the Education Code amendment, focusing exclusively on personal rights liability for social media platforms.

The civil penalty structure in Section 3273.74(a) was substantially reduced. The previous three-tier system based on months of gross revenue (intentional: 3 months, knowing/willful: 2 months, reckless: 1 month) is replaced with a two-tier fixed-dollar system: intentional/knowing/willful violations capped at $1,000,000 and reckless violations capped at $500,000. The separate 'intentional' tier (3 months gross revenue) was eliminated and merged into the knowing/willful category.

The state mandate reimbursement provision (former Section 2) requiring the Commission on State Mandates to reimburse local agencies for costs mandated by the bill was removed, consistent with eliminating the school policy requirement. The fiscal note changed from 'State-mandated local program: yes' to 'no.'

Civil penalties changed from percentage-of-revenue amounts (3 months gross revenue for intentional, 2 months for knowing/willful, 1 month for reckless) to fixed caps: up to $1,000,000 for intentional/knowing/willful violations and up to $500,000 for reckless violations. This dramatically reduces potential liability for large platforms.

Added 'liable to a prevailing plaintiff' language, clarifying that the civil penalty is awarded to the winning party in the action rather than being an unspecified remedy.

Added Civil Code Section 52.1 to the list of covered provisions (alongside Sections 51.7, 51.9, and 52), expanding which personal rights violations trigger platform liability.

Statutory sections renumbered: the definition section changed from 3273.73 to 3273.72, and the liability section changed from 3273.74 to 3273.73.

Added Assembly Members Bryan, Lowenthal, Ortega, and Pellerin as coauthors of the bill.

Added an entirely new Title 23 (Sections 3273.72-3273.74) to Part 4 of Division 3 of the Civil Code, creating a liability framework for social media platforms that violate hate crime and intimidation statutes.

The bill's overall scope expanded from a single Education Code amendment (requiring schools to adopt social media policies) to a two-part bill combining that school mandate with a new Civil Code liability regime for social media platforms.

Defined 'social media platform' for penalty purposes as one generating more than $100 million per year in gross revenues, limiting the new liability regime to large platforms.

Established a tiered civil penalty structure: up to 3 months of gross revenue for intentional violations, 2 months for knowing or willful violations, and 1 month for reckless violations. Penalties can be doubled if the platform knew or should have known the victim was a minor.

Provided that deploying an AI or algorithm that relays content to users may be considered an independent act of the platform, and that platforms are deemed to have actual knowledge of how their own AI and algorithms deliver content to different users.

Added a severability clause and an anti-waiver provision stating that any waiver of the new title is void and unenforceable as contrary to public policy.