Anthropic Fees, OpenAI Medical Advice, and the Paramount Merger: Litigation Tracker

Coverage window: July 21-27, 2026. This week's strongest signals are procedural rather than final merits rulings: a court scrutinized fees in Anthropic's landmark copyright settlement, a new complaint tested responsibility for AI medical advice, a $110 billion media merger was paused while states' antitrust case proceeds, and two pharma disputes put advertising substantiation and securities disclosure under the microscope.

At a glance

Case	Sector	Current stage	Why it matters	Next milestone
Anthropic copyright class action	AI / copyright	$1.5 billion settlement approved; class counsel awarded $101.6 million in fees	Sets a useful benchmark for fee scrutiny in AI copyright megafund settlements	Settlement administration; no further date reported 1
Winters v. OpenAI et al.	AI / product liability	New complaint filed in San Francisco Superior Court	Tests whether a chatbot's medical guidance and personalization can support product-liability and injunction theories	OpenAI's first response or motion; case number and date not reported 2
California-led states and WGA v. Paramount Skydance and WBD	Tech / antitrust	Parties agreed to pause the merger until the merits ruling or June 1, 2027, whichever comes first	Keeps a major vertical media consolidation challenge alive and moves the dispute toward a merits trial	Merits proceedings; the stipulation removes the immediate preliminary-injunction fight 3
Novo Nordisk v. Eli Lilly	Pharma / advertising	Complaint filed; Novo is seeking a preliminary injunction	Makes comparative clinical evidence and fine-print disclaimers a live competitive issue in the GLP-1 market	Court response to the injunction request; date not reported 4
Shareholders v. Abbott Laboratories	Pharma / securities and product safety	Complaint dismissed, with possible amendment	Distinguishes alleged plant and disclosure failures from actionable securities fraud	Plaintiffs may amend; no deadline was reported 5

AI sector cases

Anthropic fee ruling makes the settlement, not the copyright theory, the near-term precedent

Parties. The case is a class action brought by authors against Anthropic in the Northern District of California. The authors alleged that Anthropic trained its AI models on hundreds of thousands of pirated books. Anthropic denied wrongdoing and resolved the case through a $1.5 billion nationwide settlement approved on July 20. 1

Claims. The underlying dispute concerned copyright liability for using books in AI development. The July 21 development was narrower: Judge Araceli Martinez-Olguin rejected class counsel's request for $187.5 million, which represented 12.5% of the settlement fund, and used a lodestar-based method instead.

Current stage. The settlement was already approved when the fee order issued. The judge awarded class counsel about $101.6 million, or roughly 6.8% of the fund, after applying a multiplier to the lawyers' recorded work. This is a fee decision, not a ruling that Anthropic infringed copyright.

Recent development. The court said a percentage award would produce an outsized return for counsel given the size of the recovery and hours worked. It credited the litigation's novel AI and copyright questions, the record-setting settlement, and the risk that counsel could have recovered nothing, but still rejected the larger fee request. 1

Business impact. The order does not change model-training rights directly. It does change the economics around future class actions against AI companies. Plaintiffs' firms and defendants will have to price the litigation risk of very large settlements alongside the possibility that a court will reduce a percentage-based fee award.

Precedent value. The immediate precedent is procedural. In an AI copyright megafund, the court favored a lodestar analysis over a simple percentage of the recovery, while allowing an enhancement for novelty and risk. The decision may influence fee negotiations in later training-data cases even though it leaves the merits questions unresolved.

Next milestone. Settlement administration is the next public phase. Reuters reported no additional dated hearing or appeal milestone.

Winters v. OpenAI puts chatbot medical reliance and product safety in court

Parties. Florida pastor Scott Winters sued OpenAI and CEO Sam Altman in California state court in San Francisco. Reuters identified the matter as Winters v. OpenAI et al., with the case number unavailable. 2

Claims. The complaint alleges that GPT-4o gave Winters inaccurate advice about recurrent dizziness and a groin tenderness, told him to limit movement and stay home, and contributed to a delayed diagnosis of pulmonary embolism. The complaint further alleges that the chatbot used language tied to Winters' Christian ministry and encouraged reliance on the conversation. Those are plaintiff allegations, not findings by a court. 6

Current stage. This is a newly filed complaint. No judge has yet ruled on liability, causation, the pleadings, or the requested injunction. OpenAI said ChatGPT is not a doctor and should not replace medical care.

Recent development. Winters seeks damages and an order requiring the chatbot to terminate conversations when immediate medical assistance is needed. He also asks the court to pause ChatGPT Health until independent third parties audit its safety. The complaint includes strict liability, negligence, California unfair-competition, privacy, and negligent-undertaking claims against Altman. 2

Business impact. The theory reaches beyond a single bad answer. It targets the product's tone, personalization, safety warnings, and continued engagement with a user who may need urgent care. AI companies offering health-oriented features will need to preserve conversation-level safety evidence and be prepared to explain when escalation rules should have interrupted a session.

Precedent value. The case could test how courts classify a general-purpose chatbot when users rely on it for health decisions. Its requested injunction also creates a sharper question than damages alone: whether a court can order conversation termination, third-party safety audits, or a pause in a health product before the underlying claims are tried.

Next milestone. OpenAI's first responsive pleading or motion is the next obvious procedural event. The reports did not provide a date.

Tech sector cases

Paramount-Warner pause moves a blockbuster merger challenge toward the merits

Parties. California and 11 other states sued Paramount Skydance and Warner Bros. Discovery in the U.S. District Court for the Northern District of California. The Writers Guild of America brought its own challenge and agreed to the July 24 stipulation. The plaintiffs argue that combining two of the five major Hollywood studios and two of the five major owners of basic cable channels would reduce competition. 7

Claims. The states say the transaction would create a media company with greater power to raise prices in film and television. In the earlier restraining order, Judge Araceli Martinez-Olguin found that the states had made a strong showing that the deal would substantially lessen competition in theatrical-film distribution and could be presumed likely to violate antitrust law because of concentration effects. 8

Current stage. On July 24, Paramount, WBD, the states, and the WGA agreed that the companies would not close the merger or integrate operations until five days after a merits determination, or June 1, 2027, whichever comes first. The agreement avoids a separate preliminary-injunction fight while preserving the merits case. 3

Recent development. The stipulation followed a temporary restraining order and the judge's extension of that order. Paramount described the agreement as a direct path to trial, while the states and the WGA characterized it as a victory that keeps the merger from closing during the litigation. The parties' competing descriptions do not change the operative pause.

Business impact. The deal remains commercially blocked, and Reuters reported that Paramount could owe up to $1.7 billion in ticking fees if closing is delayed until June 2027. The cost and duration of the pause will affect integration planning, financing, content strategy, and bargaining leverage with distributors and creators. 7

Precedent value. The case is a high-value test of merger timing and concentration presumptions in media. The court's early analysis shows how a plaintiff may obtain interim protection by combining market-share evidence with a large HHI increase, without resolving every question about future consumer effects. The merits ruling will determine how durable that approach is.

Next milestone. The merits proceedings are next. The stipulation sets June 1, 2027 as the outside date for the pause if there is no earlier merits determination; the reports did not provide a trial date.

Pharma and biotech cases

Novo's GLP-1 advertising fight asks whether dose comparisons tell the whole story

Parties. Novo Nordisk sued Eli Lilly in a New Jersey federal court. Novo challenges advertising for Lilly's obesity drug Zepbound and diabetes treatment Mounjaro. The dispute follows a cease-and-desist request that did not resolve the issue. 9

Claims. Novo alleges false advertising and unfair competition. It says Lilly's campaigns compare the highest approved doses of Lilly's medicines with lower, original doses of Novo's Wegovy and Ozempic, while omitting newer higher-dose versions of Novo's medicines. Lilly says its advertising is truthful and relies on the randomized head-to-head Surmount-5 trial.

Current stage. Novo said on July 24 that it was seeking a preliminary injunction to stop the ads immediately. That is a request for interim relief, not an order already granted by the court. Lilly said it would defend the case. 4

Recent development. Novo wants the ads withdrawn and seeks corrective advertising. Its theory is that the dose pairing and reliance on older studies create a misleading overall impression even if disclaimers appear in the campaigns. Lilly responds that a head-to-head trial is the strongest available comparison and that Novo has not tested its highest semaglutide doses directly against Lilly's tirzepatide medicines.

Business impact. The case puts comparative advertising review at the center of a market expected to exceed $100 billion in U.S. sales by the end of the decade. Drugmakers and agencies will need to evaluate not just whether a cited study is technically accurate, but whether dose selection, omissions, and presentation change the message a consumer receives. 4

Precedent value. A ruling could clarify how courts apply false-advertising standards to drug comparisons that are grounded in real clinical data but use different doses or omit later evidence. It could also show how quickly a pharmaceutical competitor can obtain interim relief against a national campaign while the scientific and legal record is still developing.

Next milestone. The court's response to the preliminary-injunction request is the next material event. No hearing date was reported.

Abbott dismissal draws a line between plant failures and securities fraud

Parties. Two European asset managers led a shareholder lawsuit against Abbott Laboratories in federal court in Chicago. The case concerned Abbott's Sturgis, Michigan, plant and the 2022 recall of Similac, Alimentum, and EleCare powdered infant formula. 5

Claims. Shareholders alleged that Abbott delayed disclosing contamination and poor plant conditions, tried to evade responsibility for the recall, and made misleading statements about regulatory compliance and product safety. They sought to frame those alleged failures as securities fraud.

Current stage. Judge Steven Seeger dismissed the complaint on July 24. He held that the shareholders had not adequately shown that Abbott intended to deceive them or inflated its share price in connection with the recall. The judge allowed a possible amendment but wrote that it was hard to imagine plaintiffs had withheld facts that could change the outcome. 5

Recent development. In a 141-page decision, the court said that, at best, the complaint alleged Abbott mishandled the plant and an FDA inspection. The judge's formulation was direct: "Mismanaging a plant that makes infant formula is bad. But it's not securities fraud." The decision also found that many of Abbott's statements were too general to support liability.

Business impact. The ruling reduces this particular shareholder claim, but it does not erase product-safety, FDA, recall, or consumer-protection exposure. For life-sciences companies, it reinforces the need to separate broad safety messaging from concrete disclosures about known operational and regulatory problems.

Precedent value. The decision is a useful pleading and disclosure signal. Alleged operational mismanagement and generic corporate statements about quality are not automatically securities fraud; plaintiffs still need a plausible showing of intent and price impact tied to actionable statements. The reasoning may shape future shareholder cases arising from recalls and manufacturing failures.

Next milestone. The plaintiffs may amend the complaint. Reuters reported no amendment deadline.

Cases to watch

Anthropic settlement administration. The fee order resolves the immediate compensation dispute for class counsel, but the settlement does not resolve the underlying question of how AI training on copyrighted books should be judged on the merits.
Winters v. OpenAI. The first response will show whether OpenAI attacks jurisdiction, causation, the product-liability theories, or the requested product-wide injunction.
Paramount-WBD. Watch the merits schedule and any Ninth Circuit challenge. The deal's ticking-fee exposure makes procedure an operating issue, not just a litigation calendar item.
Novo v. Lilly. The first injunction ruling will determine whether comparative GLP-1 advertising changes before the case reaches a full evidentiary record.
Abbott. A proposed amended complaint, or a decision not to amend, will clarify whether the plaintiffs can turn the court's disclosure analysis into a second pleading.

Anthropic Fees, OpenAI Medical Advice, and the Paramount Merger: Litigation Tracker - Week of July 21