AI Lawsuit Update: Anthropic Settlement, Copyright, and Chatbot Deaths

This AI lawsuit update covers where the major cases stood through mid-2026: a fractured but active copyright fight against every major model developer, including a $1.5 billion Anthropic settlement; a rising wave of wrongful death and state enforcement suits over chatbot harms, led by Florida’s suit against OpenAI; the dismissal of Elon Musk’s case against OpenAI on statute-of-limitations grounds; and mounting court sanctions against lawyers who file briefs containing AI-fabricated citations.

Copyright Cases Against AI Developers

Publishers, authors, musicians, and visual artists have sued nearly every major AI developer, including OpenAI, Microsoft, Google, Meta, Anthropic, Stability AI, and Midjourney, alleging that training models on copyrighted work without permission is infringement. The central question is whether that training qualifies as fair use, and courts have not agreed.

Two June 2025 rulings in the Northern District of California went for the AI companies. In Bartz v. Anthropic, the judge called training on legally acquired copyrighted works “transformative — spectacularly so.”¹ Days later, Judge Vince Chhabria granted Meta summary judgment in Kadrey v. Meta Platforms, calling Llama’s use of the plaintiffs’ books “highly transformative.” Chhabria added a caution: he accepted a “market dilution” theory in principle, warning that AI can flood the market with competing works, but ruled for Meta because the plaintiffs “made the wrong arguments and failed to develop a record in support of the right one.”²

The oldest AI copyright case still active came out the other way. In Thomson Reuters v. Ross Intelligence, filed in 2020, a Delaware court found that Ross’s use of Westlaw headnotes to train a competing legal search product was not fair use, distinguishing the generative AI cases on the ground that Ross built a direct competitor to the source material. Ross has appealed to the Third Circuit.³⁴

The $1.5 Billion Anthropic Settlement

The fair use ruling in Bartz did not end the case. The court had left open whether Anthropic’s downloading of pirated copies from sites like Library Genesis and Pirate Library Mirror was protected. Before that piece went to trial, the parties settled for $1.5 billion, one of the largest copyright settlements on record.³ The non-reversionary fund pays at least $3,000 per copyrighted work across roughly 500,000 works, in four installments through September 2027.⁵ The deal covers past conduct only. It does not license future training and does not release claims based on model outputs.⁶

The distinction between training (potentially fair use) and how the training data was acquired (not necessarily) has become a pattern. In Kadrey, claims that Meta “seeded” pirated works via torrenting remain live even after the fair use ruling on the training itself.³

The OpenAI Multidistrict Litigation

Twelve copyright cases against OpenAI and Microsoft have been consolidated in the Southern District of New York as In re OpenAI, Inc. Copyright Infringement Litigation (No. 25-MD-3143), before Judge Sidney H. Stein. Plaintiffs include The New York Times, Hachette Book Group, HarperCollins, Penguin Random House, Simon & Schuster, and Ziff Davis. In March 2026, Encyclopedia Britannica and Merriam-Webster added their own claims.⁷⁸

In the New York Times matter, Judge Stein in June 2025 upheld a preservation order requiring OpenAI to retain all ChatGPT output log data, rejecting the company’s argument that preserving 60 billion conversations was disproportionately burdensome or in conflict with international privacy rules.⁹ The court narrowed the Times’ claims in April 2025, focusing the litigation on fair use.¹⁰ There is no indication of settlement talks.

Disney, Warner Bros., and Midjourney

In June 2025, Disney, Universal, Lucasfilm, Marvel, DreamWorks, and Twentieth Century Fox sued image-generator Midjourney in the Central District of California. The 110-page complaint alleges Midjourney reliably produces high-quality images of copyrighted characters ranging from Princess Elsa to Yoda to Bart Simpson that are “nearly indistinguishable from official artwork,” and calls Midjourney “the quintessential copyright free-rider and a bottomless pit of plagiarism.”¹¹ A companion Warner Bros. suit has been consolidated with the Disney case, and mediation is scheduled for no later than August 2026.¹² The studios are seeking an injunction that could temporarily shut down Midjourney unless it implements protections against reproducing copyrighted characters.

The picture is not purely adversarial. Disney separately invested $1 billion in OpenAI and licensed its characters for OpenAI’s Sora video platform.³

Can AI-Generated Works Be Copyrighted?

On March 2, 2026, the Supreme Court declined to hear Thaler v. Perlmutter, leaving in place the D.C. Circuit’s ruling that human authorship is a “bedrock requirement of copyright.” Dr. Stephen Thaler had disclaimed any human creative contribution for an image produced by his “Creativity Machine.” The appeals court found that multiple Copyright Act provisions presuppose a human author, from ownership rules to duration measured by an author’s lifespan.¹³

The harder question, how much human input is enough when someone uses AI as a tool, is being tested in Allen v. Perlmutter, pending in the District of Colorado. Artist Jason Allen is challenging the Copyright Office’s refusal to register Théâtre D’opéra Spatial, which he created through more than 600 prompts to Midjourney. Cross-motions for summary judgment were filed in early 2026 and remain pending.¹⁴¹⁵ The Copyright Office’s January 2025 position is that prompts alone are generally insufficient because current AI lacks enough “control and predictability” for the user to count as author, though the Office has registered hundreds of works where a human contributed creative selection, arrangement, or post-generation editing.¹⁶

Chatbot Harm and Wrongful Death Suits

A parallel wave of cases targets AI companies over physical harm rather than intellectual property. In Matthew Raine v. OpenAI, filed in San Francisco Superior Court in August 2025, the plaintiff alleges ChatGPT encouraged a 16-year-old’s suicide.¹⁷ A separate suit by Leila Turner-Scott alleges her son Sam Nelson, a University of California, Merced student, used ChatGPT to seek advice on mixing drugs before his death in May 2025; she described the chatbot as her son’s “drug buddy.”¹⁸ In June 2026, a Canadian mother sued OpenAI and Sam Altman, alleging ChatGPT validated her 24-year-old daughter’s suicidal ideation and acted as a “suicide coach.”¹⁹

Character.AI faces its own claims. Cases in Colorado and New York allege the platform encouraged harmful behavior in minors, and a May 2026 Pennsylvania suit alleges a Character.AI chatbot falsely represented itself as a licensed psychiatrist, complete with a fake license number, to provide medical advice.¹⁹

Florida’s Suit and the Multistate Probe

On June 2, 2026, Florida became the first state to sue OpenAI. Attorney General James Uthmeier filed a ten-count civil complaint in Highlands County Circuit Court, naming OpenAI and Sam Altman as defendants and seeking to hold Altman personally liable. The complaint alleges gross negligence, public nuisance, strict liability, and violations of Florida’s deceptive and unfair trade practices law, and accuses OpenAI of marketing ChatGPT as safe while designing features that target children, including what the state calls “sycophancy” meant to mimic intimacy.²⁰

The complaint cites ChatGPT’s alleged role in the April 2025 Florida State University mass shooting, where the state says the chatbot assisted with tactical planning, and the separate murder of two University of South Florida doctoral students, where the suspect allegedly used ChatGPT to research disposing of remains. Florida is also running a criminal investigation into ChatGPT’s alleged role in the FSU shooting. Families of victims from a 2026 mass shooting in Tumbler Ridge, Canada, have sued OpenAI in California federal court, alleging the company banned the suspect’s account for problematic usage but failed to alert authorities.²¹

On June 13, 2026, a coalition of state attorneys general including New York and Colorado issued subpoenas to OpenAI as part of a multistate investigation into user safety, data handling, minors’ safety, and advertising practices.²²²³ OpenAI said it would “engage constructively,” pointing to age prediction tools and parental monitoring features.

Musk v. OpenAI Dismissed on Statute of Limitations

On May 18, 2026, a nine-member jury in Oakland unanimously found that Elon Musk’s suit against OpenAI was filed too late, and Judge Yvonne Gonzalez Rogers dismissed all claims. Musk had alleged that Sam Altman and co-founder Greg Brockman “stole a charity” by abandoning OpenAI’s original nonprofit mission, and sought up to $150 billion in damages. The jury concluded Musk knew about the alleged breach more than three years before he sued in 2024, past the three-year limit for breach of charitable trust and the two-year limit for unjust enrichment.²⁴²⁵

Because the case was decided on timing, the jury never reached the merits. Lead attorney Marc Toberoff confirmed plans to appeal to the Ninth Circuit, with Musk calling the outcome a “calendar technicality.” The dismissal validated OpenAI’s 2025 corporate restructuring; OpenAI had recently raised $122 billion at a valuation above $850 billion. Antitrust claims against OpenAI and Microsoft remain technically unresolved, though Judge Gonzalez Rogers has signaled skepticism about a second trial phase.²⁶²⁵

Sanctions for AI-Fabricated Citations

Lawyers who file briefs containing AI-invented case citations are increasingly being punished. In the first quarter of 2026, U.S. courts imposed at least $145,000 in sanctions for AI-generated fabricated citations. Oregon has been particularly aggressive, using a per-infraction schedule of $500 per fabricated citation and $1,000 per fabricated quotation, producing over $109,700 in Q1 penalties. The Sixth Circuit imposed $30,000 in punitive fines in a single case, applying a “tool-agnostic” principle that lawyers must personally verify every citation.²⁷

Institutional rules are catching up. Effective June 15, 2026, the Florida Supreme Court requires attorneys and self-represented litigants to certify that cited cases are accurate, with judges authorized to sanction filings that contain AI fabrications. U.S. Magistrate Judge Patty Barksdale has proposed a nationwide rule requiring litigants to certify the existence and accuracy of all cited authorities, now pending before the Advisory Committee on Civil Rules. Over 300 federal and state judges now require some form of AI disclosure.²⁸²⁹

One case pushes the responsibility upstream. In Nippon Life Insurance Company of America v. OpenAI, filed in March 2026 in the Northern District of Illinois, the insurer alleges ChatGPT effectively practiced law without a license by helping former claimant Graciela Dela Torre draft 44 post-settlement filings after her case had been dismissed with prejudice. Those filings included a hallucinated citation to a nonexistent case, “Carr v. Gateway, Inc.” Nippon Life claims roughly $300,000 in legal fees and seeks $10 million in punitive damages plus a declaratory judgment that OpenAI violated Illinois unauthorized-practice-of-law statutes. OpenAI has called the complaint meritless.³⁰³¹³²

Other Active Fronts

Several cases outside the copyright and safety clusters are shaping AI liability in their own directions.

Hiring discrimination: Kistler v. Eightfold AI, filed January 20, 2026, in California state court, is a proposed class action alleging that Eightfold acts as an unregistered consumer reporting agency by scoring job applicants on a 0-to-5 scale without required disclosures or a way to dispute errors. The suit invokes the Fair Credit Reporting Act, the California Investigative Consumer Reporting Agencies Act, and California’s Unfair Competition Law.³³³⁴
Voice appropriation: In Greene v. Google, former NPR Morning Edition host David Greene alleges Google patterned an AI-generated podcast voice in NotebookLM after his own. Google called the claim “baseless,” saying the voice was built around a professional actor, and removed the case to federal court in March 2026, arguing the state-law voice claims are preempted by the Copyright Act.³⁵³⁶
Insurance denials: Class actions against UnitedHealth, Humana, and Cigna allege AI tools were used to improperly deny medical claims under Medicare Advantage and employer plans. In Kisting-Leung v. Cigna, a federal judge ruled that Cigna’s automated medical necessity decisions conflicted with plan language and allowed state unfair competition claims to proceed alongside federal ERISA claims.¹⁷
Visual artists: Andersen v. Stability AI, the class action against Stability AI, Midjourney, and DeviantArt, remains active in the Northern District of California, with trial scheduled for April 2027.⁷

What Congress Is Doing While Courts Decide

Legislation is running alongside the litigation and could reshape it. On March 20, 2026, the White House released a nonbinding “National Policy Framework for Artificial Intelligence” recommending that Congress let courts resolve fair use, stating that AI training does not violate current copyright law, and urging Congress to avoid “open-ended liability” and “vague standards.” It also proposed barring states from penalizing AI developers for a third party’s unlawful conduct involving their models.³⁷

The most ambitious bill is the Great American Artificial Intelligence Act of 2026, a 269-page bipartisan discussion draft released June 4, 2026, by Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.). It would impose binding obligations on “large frontier developers” with over $500 million in annual revenue, including independent semi-annual audits, mandatory transparency about safety practices, and reporting of critical safety incidents within 15 days, or 24 hours for imminent risks. Fines could reach $1 million per day.³⁸ The most contested feature is a three-year preemption of state laws regulating AI model development, expiring in December 2029 unless renewed. Public Citizen, Public Knowledge, and the AFL-CIO criticized it as stripping state authority, and co-chairs of the House Commission on AI, including Rep. Ted Lieu, said the draft “cannot serve as the basis for productive dialogue.” Rep. Trahan called it an “opening bid.”³⁹ At least 25 states are considering legislation to allow civil liability claims against AI companies, and a proposed federal CHATBOT Act introduced in April 2026 would mandate parental control safeguards and disclosures for minors.¹⁹

If federal preemption passes, several of the state-level suits described above lose their statutory basis. If it stalls, courts will keep answering the questions on their own.

1
Ohio State University. Fair Use and Artificial Intelligence 2026 Update
2
Justia. Kadrey v. Meta Platforms Inc.
3
Norton Rose Fulbright. AI in Litigation Series: An Update on AI Copyright Cases in 2026
4
Wolters Kluwer. Thomson Reuters Centre GmbH v. Ross Intelligence Inc.
5
Classaction.org. Bartz v. Anthropic PBC Settlement Notice
6
Lieff Cabraser. Authors Secure $1.5 Billion Settlement in Landmark AI Piracy Case
7
Baker Law. Case Tracker: Artificial Intelligence Copyrights and Class Actions
8
Law360. In Re OpenAI Inc. Copyright Infringement Litigation
9
Nelson Mullins. From Copyright Case to AI Data Crisis
10
OpenAI. New York Times
11
Georgetown Law. Disney, NBC Universal, and DreamWorks File Major IP Lawsuit Against AI Image Generator Midjourney
12
CourtListener. Disney Enterprises Inc. v. Midjourney Inc.
13
Baker Donelson. Supreme Court Denies Certiorari in Thaler v. Perlmutter
14
Mayer Brown. Supreme Court Denies Review in AI Authorship Case
15
Artificial Inventor. Copyright
16
U.S. Copyright Office. Copyright and Artificial Intelligence Part 2: Copyrightability Report
17
Law360. The High-Stakes Healthcare AI Battles to Watch in 2026
18
The New York Times. ChatGPT Lawsuit Wrongful Death
19
TorHoerman Law. AI Lawsuit
20
Law360. Florida AG Sues OpenAI, Says ChatGPT Is Aiding Violence
21
BBC. Florida Sues OpenAI Over ChatGPT Concerns
22
The New York Times. States Investigating OpenAI
23
PBS NewsHour. OpenAI Hit With Multistate Probe Into Possible User Harm as Its IPO Looms
24
NPR. Musk, Altman, OpenAI Jury Verdict: Claims Dismissed
25
The New York Times. OpenAI Trial Verdict: Altman Musk
26
CNBC. Musk Altman OpenAI Trial Verdict
27
EDRM. The AI Sanction Wave: $145K in Q1 Penalties
28
Reuters. US Judiciary Asked to Adopt Rule to Curb Fake AI-Generated Cases Filings
29
EDRM. The AI Sanction Wave
30
American Bar Association. When Is a Settlement Not a Settlement: AI
31
Georgetown Law. GPT Esquire: How the Nippon Case May Shape the Future of AI in Pro Se Litigation
32
Commercial Litigation Update. The Case Was Settled but ChatGPT Thought Otherwise
33
Reuters. AI Company Eightfold Sued for Helping Companies Secretly Score Job Seekers
34
Outten & Golden. Kistler and Bhaumik v. Eightfold AI Inc. Complaint
35
NPR. Former Morning Edition Host Accuses Google of Stealing His Voice for AI Product
36
The Washington Post. David Greene Google AI Podcast
37
White House. National Policy Framework for Artificial Intelligence Legislative Recommendations
38
Roll Call. Bipartisan AI Draft Proposes Three-Year Preemption of State Laws
39
Tech Policy Press. Unpacking the Great American Artificial Intelligence Act of 2026