News April 03 2026

April 3, 2026 — Daily digest of AI law developments.

This article consolidates 3 news stories from April 3, 2026.

Contents

1. California SB 813 AI Safety Commission 2. Ted Entertainment, Inc. 3. Tennessee AI Personhood Bill SB 837

---

California SB 813 AI Safety Commission

California SB 813 would establish the California AI Standards and Safety Commission within the state Government Operations Agency to oversee AI safety regulation, including designating independent verification organizations and coordinating AI risk assessment across state agencies.

Bill Provisions

The bill adds Chapter 14, commencing with Section 8898, to Division 1 of Title 2 of the California Government Code. Key provisions include:

• Commission establishment: Creates the California AI Standards and Safety Commission within the Government Operations Agency
• Liaison relationships: Requires the Commission to maintain formal liaison relationships with state agencies deploying or procuring AI
• Technical expertise: The Commission would provide AI technical expertise and risk assessments
• Independent Verification Organizations (IVOs): Authorizes the Commission to designate one or more IVOs, requiring them to implement AI risk mitigation plans covering cybersecurity, chemical/biological/radiological/nuclear threats, malign persuasion, and model autonomy/exfiltration
• Reporting: IVOs must submit annual reports to the Legislature and Commission on evaluation resources, mitigation adequacy, and risks
• Fee authority: The Commission may establish fees for applicants and IVOs to offset costs
• Multistakeholder Regulatory Organizations (MROs): The Attorney General would designate MROs for AI model and application certification, decertification, and reporting

Legislative History

• February 21, 2025: Bill introduced by Senator McNerney
• May 12, 2025: Placed on Senate Appropriations suspense file
• January 5, 2026: Author's amendments read; re-referred to Senate Appropriations Committee
• January 22, 2026: Senate Appropriations approved (5-2, "Do pass")
• January 27, 2026: Passed the Senate 31-7; ordered to the Assembly
• As of April 13, 2026: Held at Desk in the Assembly, pending referral

Current Status

SB 813 has passed the Senate and is pending in the Assembly, where it awaits committee referral. The Transparency Coalition's April 10, 2026 update noted the bill was "revived and re-referred to the Senate Appropriations Committee" after earlier passage, though the bill had actually already cleared Appropriations and the full Senate; it is now awaiting Assembly action.

Significance

If enacted, SB 813 would create one of the most comprehensive state-level AI regulatory bodies in the United States, with authority over AI safety verification, risk assessment, and coordination with other state agencies. The bill's IVO framework mirrors concepts from the California SB 1047 debate but focuses on a commission-based oversight model rather than direct developer liability.

References

See individual article: California SB 813 AI Safety Commission

---

Ted Entertainment, Inc.

On April 3, 2026, Ted Entertainment, Inc. filed a trio of class-action lawsuits in federal court against OpenAI, Apple, and Amazon, alleging violations of the Digital Millennium Copyright Act (DMCA) by circumventing YouTube's technological protection measures to scrape copyrighted videos for AI training.

The Lawsuits

The cases—Ted Entertainment v. OpenAI Inc. (N.D. Cal., Case 3:26-cv-02935), Ted Entertainment v. Apple Inc. (N.D. Cal., Case 3:26-cv-02936), and Ted Entertainment v. Amazon.com, Inc. (W.D. Wash., Case 2:26-cv-01134)—were filed by plaintiffs Ted Entertainment (creator of "h3h3 Productions" and "H3 Podcast Highlights"), Matt Fisher ("MrShortGame Golf"), and Golfholics.

Collectively, the plaintiffs operate YouTube channels with over 6 million subscribers and 4.3 billion views.

Key Allegations

The plaintiffs allege defendants bypassed YouTube's technological protection measures—including streaming-only delivery, API limits, access controls, rolling ciphers, and IP detection systems—to download millions of videos at scale.

OpenAI and Sora

The complaint against OpenAI targets its Sora text-to-video AI model. Plaintiffs identified their content in multiple datasets including HD-VILA-100M (146-285 videos), HD-VG-130M (83-209 videos), and Panda-70M (155-283 videos).

Apple and Panda-70M

The lawsuit against Apple targets "Apple AI Video" models allegedly trained using the Panda-70M dataset—70 million clips split from YouTube videos released by Snap in 2024.

Amazon and Nova Reel

The suit against Amazon targets Nova Reel, Amazon's text-to-video AI model accessible via AWS Bedrock.

Legal Theory

All three cases allege violations of 17 U.S.C. § 1201(a), the DMCA's anti-circumvention provision, which prohibits circumventing technological measures that control access to copyrighted works. Plaintiffs seek statutory damages of up to $150,000 per violation, class certification, and permanent injunctions.

Significance

These cases mark the seventh lawsuit filed by Ted Entertainment against AI companies, following prior actions against Meta, Nvidia, ByteDance, and Snap. The litigation tests whether downloading YouTube videos for AI training—even when content is publicly viewable—violates the DMCA when TPMs are circumvented.

References

See individual article: Ted Entertainment, Inc.

---

Tennessee AI Personhood Bill SB 837

The Tennessee AI Personhood Bill (SB 837 / HB 849) is legislation in the Tennessee General Assembly that defines "human being," "life," and "natural person" for statutory construction purposes, explicitly excluding artificial intelligence systems, computer algorithms, software programs, computer hardware, and machines from these definitions.

Bill Provisions

The bills amend Tennessee Code Annotated, Title 1, relative to conditions of personhood. The core provision states:

As introduced, defines "human being," "life," and "natural person" for statutory construction purposes; excludes from the definition of "person," "life," and "natural person" artificial intelligence, a computer algorithm, a software program, computer hardware, or any type of machine.

This language directly clarifies that AI is not considered a person, life, human being, or natural person under state law, preventing any legal recognition of machine-based entities as persons.

Legislative History

SB 837 passed the full Tennessee Senate on April 6, 2026 with a 26-6 vote. The House companion HB 849 was set aside in favor of SB 837, which then passed the full House on April 8, 2026 with a 93-2 vote. The bill was sent to Governor Bill Lee on April 15, 2026.

Current Status

As of April 26, 2026, SB 837 has passed both chambers and has been delivered to Governor Lee's desk for signature or veto. The governor has 10 days to act on the bill. If signed, the law would take effect on the date specified in the legislation.

Significance

This legislation represents one of the first state-level efforts to formally exclude AI from legal personhood definitions, addressing concerns that AI systems might someday claim legal rights or be recognized as legal persons. If enacted, it would establish a statutory construction principle that prevents courts or agencies from interpreting "person" to include AI systems in Tennessee law.

Related Legislation

• Tennessee CHAT Act (SB 1700) — chatbot safety and data privacy
• Tennessee Healthcare AI Bill (SB 1580) — prohibits AI from claiming to be a mental health professional (signed into law April 1, 2026)

References

See individual article: Tennessee AI Personhood Bill SB 837

---