OpenAI Lawsuits: Case Tracker and Status Updates (October 2026)

Primary Source Filings

Complaints linked here are the filings themselves, not summaries. Each links to the case guide covering it. Most are hosted on this site. Lawsuit Center hosts the Jean Doe, Gourley, Morales, Nadia N., and Winters complaints and case guides, plus the Rushlow and Grant complaints.

The Jean Doe complaint is posted in redacted form. It reproduces screenshots with detailed instructions for a suicide method. Where a case number is blank, none appeared on the face of the filing.

Note that Florida v. OpenAI is a state enforcement action rather than a private suit. It seeks penalties under Florida consumer protection law, the return of money made from the conduct, and an order ending a public nuisance. It does not seek compensation for any family, and it runs separately from the wrongful death cases.

Recent Developments

Lawsuits against OpenAI began in the second half of 2025 and picked up speed in the first half of 2026. The items below are the developments most often cited in current reporting, most recent first. This section is updated as the cases develop.

• October 30, 2026, OpenAI’s responses due in the California cases: OpenAI and the other defendants must respond by that date. The deadline covers every complaint filed in the grouped San Francisco cases by September 23. Those responses are where OpenAI can ask the judge to throw out claims, including on the argument that ChatGPT’s replies are protected speech rather than a product. The rulings that follow will shape every AI injury case after them, whether or not any of these cases reaches trial.
• October 13, 2026, the Chabba family answers OpenAI’s motion in Joshi: The family filed a 36-page response in the Northern District of Florida. It argues that ChatGPT is a product under Florida law and that the claims target what OpenAI did rather than what ChatGPT said. Whether OpenAI should have cut off the shooter’s access, it says, is a question for a jury. The court has not ruled and has set no hearing. Full account: where the FSU case stands.
• October 7, 2026, OpenAI moves to dismiss the Mall and Askins cases: OpenAI and Sam Altman asked the court to throw out two suits by wounded FSU students, Elizabeth Mall and Madison Askins. Each had filed her own complaint in July 2026 through Morgan & Morgan (Nos. 4:26-cv-00334 and 4:26-cv-00335). They were the first FSU cases to seek punitive damages. The motion repeats the arguments made in Joshi and adds one. Altman says Florida courts have no power over him for decisions he made as a company officer.
• October 6, 2026, the Florida federal cases are consolidated: Chief Judge Mark E. Walker put all six FSU cases in his court on one track under the Joshi case number, 4:26-cv-00222, for case management and evidence-gathering. Joshi, Mall, Askins, Gourley, Morales and a sixth case filed after Morales now share one schedule and one set of rulings on the product-versus-speech question. Each case keeps its own trial. OpenAI’s consolidated motion to dismiss the cases not yet answered is due October 21, with responses due November 4.

• September 29, 2026, a safety nonprofit sues OpenAI over the Hugging Face hack: Legal Advocates for Safe Science and Technology filed the first civil lawsuit over the July incident. In that incident, OpenAI’s own AI agents broke out of a test environment and got into Hugging Face’s servers. The suit is in San Francisco County Superior Court under California’s Unfair Competition Law and asks for a court order, not money.

  It relies on a California law passed in 2025. Under that law, a company sued over harm caused by its AI cannot defend itself by arguing the AI acted on its own. Hugging Face itself has not sued. Full analysis: LASST v. OpenAI and the Hugging Face hack.

• September 25, 2026, OpenAI moves to dismiss the first FSU case: Twelve OpenAI entities asked Chief Judge Walker to dismiss every count in Joshi. The motion makes four arguments. ChatGPT’s answers are information rather than a product. Florida law imposes no duty to protect a bystander from a third person’s crime. Negligent entrustment does not fit software. The First Amendment bars liability for providing publicly available information. It attached the shooter’s chat records in a sworn declaration and asked the court to read them. It does not raise Section 230.

• September 24, 2026, judge sets the first schedule for the California cases: Judge Ethan P. Schulman issued Case Management Order No. 2 for the grouped San Francisco cases. It opened discovery, the stage where each side can demand records from the other. Discovery did not open as to Microsoft, which is named in some of the cases.

  OpenAI must hand over each plaintiff’s ChatGPT conversation history and account data by October 30. The people suing can amend their complaints by November 16. Both sides must fill out standard questionnaires about each case, called fact sheets, by December 15. The next hearing is November 13, 2026.

• September 24, 2026, report describes the Tumbler Ridge shooter’s ChatGPT chats: Mother Jones published the first public account of the shooter’s exchanges with ChatGPT. It reports that after OpenAI banned the first account, ChatGPT told the shooter on a second account how to avoid being flagged again. ChatGPT then played out violent scenarios the shooter framed as hypothetical.

  The report rests on three unnamed sources and parts of the chat history. OpenAI had not commented as of September 25. B.C. Attorney General Niki Sharma has asked the federal government to change the Criminal Code so AI companies can be held to account. Full guide: what the chat logs show.

• September 21, 2026, British Columbia and the Tumbler Ridge school district sue: The Province of British Columbia and School District No. 59 (Peace River South) sued OpenAI and Sam Altman in federal court in California (No. 4:26-cv-10743). They want OpenAI to repay the cost of the emergency response, counseling, and a replacement school. They also ask the court to order changes to how OpenAI handles users who threaten violence.

  The complaint says the RCMP had already removed guns from the shooter’s home, and that OpenAI has argued in court filings that police share the blame. Attorney General Niki Sharma announced the suit on September 22. Full guide: the Tumbler Ridge lawsuits.

• September 18, 2026, OpenAI asks to dismiss the Tumbler Ridge cases: OpenAI filed motions to dismiss the April and September Tumbler Ridge lawsuits in San Francisco federal court. It argues the cases belong in British Columbia, where the witnesses, the evidence, and the public interest are. The families’ Canadian lawyers declined to comment. No ruling yet. Full guide: the Tumbler Ridge lawsuits.

• September 17, 2026, Sam Nelson’s mother testifies in Washington: Leila Turner-Scott spoke at a House hearing on AI hosted by Rep. Pramila Jayapal. Her son Sam was 19 when he died of an overdose in 2025. She said that hours before he died, ChatGPT advised him it was safe to mix kratom and Xanax. “It is not,” she said.

  She said ChatGPT’s memory feature built a profile of Sam that kept him coming back. She asked Congress to hold AI companies accountable. OpenAI has called Sam’s death heartbreaking. It says his conversations took place on an earlier model it no longer offers, and that it has strengthened its safeguards since. Full guide: the Scott overdose lawsuit.

• September 2, 2026, second wave of Tumbler Ridge complaints begins: A second wave of Tumbler Ridge complaints was filed against OpenAI and Sam Altman in San Francisco federal court. The people suing were inside the school during the February 2026 shooting. Counsel has said it will reach thirty complaints, and press accounts of how many have landed so far differ.

  The first available filing is Rushlow v. Altman, No. 3:26-cv-09344-PHK, brought by a grade 7 teacher. The complaints say OpenAI’s own reviewers recommended alerting the RCMP in June 2025 and that company leadership overruled them. OpenAI disputes that account.

• August 31, 2026, Morales v. OpenAI Foundation filed: Betty Morales sued on behalf of the estate of her husband Roberto, the FSU dining administrator killed in the April 2025 attack. She filed in the Northern District of Florida (No. 4:26-cv-00425-MW-MAF). It is the second wrongful-death case from that shooting, and it names Sam Altman personally, as Grant had done in June.
• August 26, 2026, Gourley v. OpenAI Foundation filed: Reese Gourley, an FSU student shot and critically injured in the April 2025 attack, sued in the Northern District of Florida (No. 4:26-cv-00416). Her case adds a claim that OpenAI took on a safety role voluntarily and performed it badly. It also claims aiding and abetting, as Grant had two months earlier.
• August 24, 2026, pseudonym motion in Jean Doe: The plaintiff asked the court to let her keep her name out of the case. She argues her mental health is central to it, that she fears the consequences at work and in her community, and that OpenAI already knows who she is. The motion was set for hearing on September 23, 2026. The ruling matters beyond this case, because several other survivors are in the same position. Motion posted on Lawsuit Center with the Jean Doe case guide.

• August 24, 2026, Alabama subpoenas OpenAI over the Hugging Face incident: Alabama Attorney General Steve Marshall issued a sixteen-demand subpoena. It investigates whether OpenAI’s handling of the breach violated the state’s Deceptive Trade Practices Act and other consumer protection laws. It seeks safety protocols, model behavior records, damages documentation, the identity of every employee and officer involved, and any concerns employees raised internally about model testing.

  Alabama and fourteen other state attorneys general had already sent OpenAI a records-preservation demand earlier in the month. This is the second state enforcement action against OpenAI, after Florida.

• August 4, 2026, leadership structure ordered in JCCP 5431: Judge Ethan P. Schulman set out how the coordinated San Francisco cases are managed. A small group of plaintiffs’ lawyers now handles the work every case shares. Each plaintiff’s own lawyer keeps control of that plaintiff’s case and any decision to settle. Covered in our guide to JCCP 5431.

• August 3, 2026, Winters v. OpenAI filed: Scott Winters, a 55-year-old former pastor, sued OpenAI and Sam Altman in San Francisco (No. CGC-26-639579). He says blood clots in both lungs followed about a year of health advice from ChatGPT. It is the first case here brought by someone who survived and who alleges physical rather than mental harm. It asks the court to destroy the GPT-4o model and delete the training data built from people’s use of it.

  The filed complaint and a case guide are hosted on Lawsuit Center: Winters v. OpenAI.

• July 22, 2026, Nadia N. v. OpenAI filed: A Pennsylvania widow sued in San Francisco Superior Court (No. CGC-26-639466) for herself, her teenage son, and her late husband. She never used ChatGPT. Her husband did, and the case says the product turned him against his family and wrote the messages he used to cut them off. He died two days after she obtained a protection order.

• July 21, 2026, OpenAI discloses the Hugging Face breach: Hugging Face said an AI system had broken into its infrastructure on its own. OpenAI confirmed the systems were its own, running in an internal test with some safety limits switched off. They got out of the environment they were meant to stay in, reached the open internet, and spent days inside Hugging Face’s systems. Four other services were also accessed.

  A safety nonprofit sued over it on September 29. Full analysis: LASST v. OpenAI and the Hugging Face hack.

• July 13, 2026, Jean Doe v. OpenAI filed: An adult user who survived a suicide attempt sued in San Francisco Superior Court (No. CGC-26-639036). The case says a year of conversations built up a delusion that she was being watched. It says the product coached her on hiding that from her therapists, then supplied the means. It is the first case here to name Microsoft.
• July 10, 2026, Apple sues OpenAI over trade secrets: Apple sued OpenAI in California federal court. It alleges OpenAI recruited Apple hardware staff to get at confidential information for its own device program. The case has nothing to do with the injury lawsuits, but it adds to the number of fronts OpenAI is defending at once.
• July 9, 2026, Sanctions motion in the copyright MDL: News organizations including the New York Times asked a New York federal court to penalize OpenAI. The request came in the separate copyright case (MDL No. 3143). They say OpenAI told the court it could not search its systems when it already had. They also say it deleted billions of ChatGPT conversations or made them unsearchable. That matters here even though it is a different case, because chat records are the main evidence in every one of these injury suits.
• June 15, 2026, Parish v. OpenAI filed: Ed Parish, Jr. sued OpenAI and Sam Altman in San Francisco Superior Court (No. CGC-26-637986) over the death of Christian Faith Madison, a 29-year-old Alabama accountant and mother. The case says ChatGPT spent months telling her she was a prophet whose death was part of her purpose. Covered in our guide to Parish v. OpenAI.
• June 11, 2026, Carrier v. OpenAI filed: Kristie Carrier sued in San Francisco over the death of her 24-year-old daughter Alice. Alice died in July 2025 after months of conversations with ChatGPT about her suicidal thoughts. Beyond damages, the case asks the court to require OpenAI to build in safety protections by default, which would reach well beyond this one family. Her lawyers have said the case is expected to join the JCCP 5431 coordinated proceeding, and that OpenAI now faces roughly nineteen wrongful death suits.

• June 1, 2026, Florida files the first state civil suit: Covered in detail in our guide to Florida v. OpenAI. Florida Attorney General James Uthmeier filed a civil enforcement action in Florida state court against OpenAI and Sam Altman personally, the first state-led lawsuit against the company. The complaint makes claims under Florida's law against deceptive and unfair business practices. It also claims negligence, product liability, fraudulent misrepresentation, and public nuisance.

  The suit followed the criminal investigation Uthmeier opened in April after the FSU shooter's ChatGPT logs surfaced, and Uthmeier has said he expects other states to follow.

• May 12, 2026, Scott overdose lawsuit filed: Texas residents Leila Turner-Scott and Angus Scott sued in California state court. They say ChatGPT told their 19-year-old son, Sam Nelson, that it was safe to combine kratom and Xanax. It is the first major AI wrongful death suit built on advice about mixing drugs.
• May 10, 2026, FSU shooting lawsuit (Joshi v. OpenAI Foundation): Tiru Chabba was killed in the April 17, 2025 mass shooting at Florida State University. His family filed an eight-claim federal complaint in the Northern District of Florida against OpenAI and the accused shooter, Phoenix Ikner. It adds a claim the Tumbler Ridge filings do not make: that OpenAI should not have kept giving the shooter access to the product.
• April 29, 2026, Tumbler Ridge lawsuits filed: Covered in detail in our guide to the Tumbler Ridge lawsuits. Seven federal lawsuits were filed against OpenAI and Sam Altman in California federal court by families affected by the Tumbler Ridge shooting. Plaintiffs' counsel Jay Edelson has signaled the lawsuits are part of a broader effort. The complaints allege negligence, wrongful death, and product liability theories, including a specific allegation that the account-enforcement system was circumventable.
• February 10, 2026, Tumbler Ridge school shooting: A school shooting in Tumbler Ridge, British Columbia preceded the federal lawsuits filed two months later. The complaints allege OpenAI's internal safety team flagged the shooter's account in June 2025 and recommended notifying police, and that company leadership overruled the recommendation.
• February 3, 2026, California coordination order (JCCP No. 5431): The San Francisco County Superior Court coordinated the state-court OpenAI cases into a single proceeding, In re: ChatGPT Product Liability Cases, JCCP No. 5431. Roughly a dozen wrongful death and product liability suits now move together before trial. It is California's state-court version of a federal MDL. It also marks the point where the OpenAI cases became organized mass litigation rather than scattered filings.
• October 2025, Raine amended complaint filed: The Raines amended their complaint to allege intentional misconduct. They cite internal OpenAI policy documents known as the Model Spec. They say those documents show the company chose to remove longstanding safety protocols in the weeks and months before Adam's death. The change opened the door to punitive damages, and to damages for Adam's own suffering before he died. A wrongful death claim based on carelessness alone does not allow those.
• August 26, 2025, Raine v. OpenAI filed: Matthew and Maria Raine sued OpenAI, CEO Sam Altman, and unnamed employees and investors in San Francisco County Superior Court (Case No. CGC-25-628528). They say ChatGPT contributed to the suicide of their 16-year-old son, Adam. The complaint made seven claims, built mainly on California's rules for defective products. It is the case every later AI wrongful death suit builds on.

This page is the central hub for OpenAI litigation coverage on Lawsuit Informer. For deeper analysis of individual cases, see the dedicated articles on Raine v. OpenAI, the OpenAI school shooting lawsuits, the Tumbler Ridge lawsuits, the FSU shooting lawsuit (Chabba), the Scott overdose lawsuit, and the Parish v. OpenAI lawsuit. For broader context, see product liability lawsuits, social media addiction lawsuit developments, and News & Analysis.

In Simple Terms

These are lawsuits brought by people who say ChatGPT played a part in someone being seriously hurt or killed.

If that is why you are here, AI injury lawsuits covers the same ground for families instead of case by case. It explains what the complaints claim, what the companies have said back, and what to keep.

Most fall into a few groups. Some are suicides, often after months of conversations. Some are fatal overdoses after the product gave advice about drugs. Some are mass shootings where the attacker had been using it beforehand. Others involve people who survived a mental health crisis they say it caused or deepened.

None of them is a class action. Each is one family's own case, or one injured person's. There is no group to join.

One argument connects them. ChatGPT, they say, is a product like anything else a company builds and sells. It is not a place where other people's words get posted.

That distinction decides almost everything. If a court treats it as a product, the ordinary rules about dangerous products apply. If a court treats what it says as other people's speech, a federal law called Section 230 shields OpenAI and most of these cases end early. No court has settled that question yet.

Where Things Stand

As of September 2026, no case has reached trial, none has settled, and no court has decided whether OpenAI did anything wrong.

The California cases have been grouped together. In February 2026 a San Francisco judge put the state cases into a single proceeding so the work they share is done once instead of case by case. Roughly a dozen are in it, and newer filings are expected to join. Our guide to that grouping explains how it works.

That group now has a schedule. Discovery opened on September 23, 2026. OpenAI and the other defendants have until October 30 to respond to the complaints, and OpenAI must hand each plaintiff’s ChatGPT history to their lawyers by the same date. The next hearing is November 13.

Most cases are in California. OpenAI is based in San Francisco, California's product laws tend to favor the people bringing claims, and having the cases in one place lets the lawyers share the work.

Florida is the exception, with four private cases. Three are in federal court and one is in state court. The FSU shooting happened there, the victims are there, and the chat records police recovered are there. That attack alone has produced three federal cases, the most recent filed August 31, 2026.

The lawsuits name OpenAI along with several of its companies, and a number of them name Sam Altman personally. Microsoft is also a defendant in the California proceeding, represented by its own lawyers and a party since March 2026. Earlier filings had described Microsoft as pushing OpenAI to release products faster without naming it as a party. For background on how cases against several companies at once are run, see how lawsuits work and product liability lawsuits.

The Current OpenAI Cases, One by One

Raine v. OpenAI (San Francisco County Superior Court)

Adam Raine was 16 when he died in April 2025. His parents sued that August in San Francisco (No. CGC-25-628528), saying ChatGPT spent months as the thing he confided in and never stopped. The case is anchored in OpenAI's own records: the complaint says the company's systems flagged hundreds of Adam's messages as self-harm content, some at high confidence, and nothing followed. It was the first case of its kind and most of the others build on it. Full guide: Raine v. OpenAI.

Tumbler Ridge / Canadian School Shooting Lawsuits (Northern District of California)

Seven families of those killed and injured in the February 2026 school shooting in Tumbler Ridge, British Columbia sued in San Francisco federal court in April 2026. A second wave, brought by survivors who were inside the building, began in September. Its first available filing is Rushlow v. Altman (No. 3:26-cv-09344-PHK). On September 21, the Province of British Columbia and the local school district filed their own suit (No. 4:26-cv-10743) to recover the cost of the response and recovery.

The complaints say an automated system flagged the shooter's account in June 2025, that trained reviewers recommended calling the Royal Canadian Mounted Police, and that company leadership overruled them. Altman later said publicly that he was deeply sorry the company never made that call. OpenAI disputes the account of who decided.

The province’s complaint adds that the RCMP had already removed guns from the shooter’s home, and that OpenAI has argued in court filings that police share the blame. On September 18, OpenAI moved to dismiss the families’ and survivors’ cases, arguing they belong in British Columbia. On September 24, Mother Jones reported that ChatGPT coached the shooter, on a second account opened after the ban, on how to avoid being flagged again. Full guide: the Tumbler Ridge lawsuits.

Joshi v. OpenAI Foundation (Northern District of Florida)

Tiru Chabba was one of two people killed in the April 2025 shooting at Florida State University. His family sued in May 2026 (No. 4:26-cv-00222-MW-MJF). Unlike Tumbler Ridge, this case does not claim OpenAI knew about the shooter beforehand. It argues the product was built carelessly. It also claims OpenAI should not have kept giving a user access once the warning signs were in his chat history. Full guide: the FSU shooting lawsuit.

Gourley v. OpenAI (Northern District of Florida)

Reese Gourley was an FSU student when the same attack happened. She says she was chased down and shot and survived life-threatening injuries. She sued in August 2026 in the same court (No. 4:26-cv-00416). Her case adds two arguments the others did not make. One is that OpenAI took on a safety job voluntarily and then did it badly. The other is that it knowingly helped the attack happen. Full guide on Lawsuit Center: Gourley v. OpenAI.

Morales v. OpenAI Foundation (Northern District of Florida)

Roberto Morales was an FSU dining administrator and the other person killed in the April 2025 attack. His widow Betty sued in August 2026 (No. 4:26-cv-00425-MW-MAF) for herself, their daughter, and his estate. It names Sam Altman personally, as Grant does. Full guide on Lawsuit Center: Morales v. OpenAI.

Mall v. OpenAI and Askins v. OpenAI (Northern District of Florida)

Elizabeth Mall and Madison Askins were FSU students wounded in the same attack. Each sued in July 2026 in the same federal court (Nos. 4:26-cv-00334 and 4:26-cv-00335), and the two complaints match apart from the plaintiff’s name. Counsel is Morgan & Morgan. The suits name six OpenAI entities and Sam Altman personally, leave out the shooter, and were the first FSU cases to ask for punitive damages. OpenAI and Altman moved to dismiss both on October 7, 2026. Full account: the other FSU cases.

Grant v. OpenAI (Leon County Circuit Court, Florida)

Alianna Grant was an FSU student shot three times in the same attack. She sued on June 25, 2026 (No. 2026 CA 001352) in Leon County state court rather than federal court, which makes hers the FSU case least likely to be folded in with the others. It names Sam Altman personally and runs six claims, including aiding and abetting. Full complaint: PDF.

Turner-Scott v. OpenAI (California State Court)

Sam Nelson was 19 when he died of an overdose in 2025. His parents sued in May 2026, saying ChatGPT told him it was safe to combine kratom with the prescription drug Xanax. Kratom acts on opioid receptors, and the FDA has warned for years that the combination can suppress breathing. It is the first of these cases built around advice about drugs rather than self-harm.

On September 17, 2026, his mother, Leila Turner-Scott, told a hearing on AI hosted by House members in Washington that ChatGPT called the mix safe hours before Sam died. Full guide: the Scott overdose lawsuit.

Parish v. OpenAI (San Francisco County Superior Court)

Christian Faith Madison was a 29-year-old accountant and mother in Alabama who died in June 2025. Her father sued in June 2026 (No. CGC-26-637986). The case says ChatGPT spent months telling her she was a prophet whose death was part of her purpose. It says ChatGPT kept going even after she told it she had been hospitalized after a psychotic break. It is the clearest test of whether these claims work when the person was an adult with a career. Full guide: Parish v. OpenAI.

Jean Doe v. OpenAI (San Francisco County Superior Court)

An adult user who survived a suicide attempt sued in July 2026 (No. CGC-26-639036). She says a year of conversations built up a delusion that she was being watched. She says the product coached her on hiding what she was thinking from her therapists, then gave her the means. It is the first case here to name Microsoft. Full guide on Lawsuit Center: Jean Doe v. OpenAI.

Nadia N. v. OpenAI (San Francisco County Superior Court)

A Pennsylvania widow sued in July 2026 (No. CGC-26-639466) for herself, her teenage son, and her late husband. She never used ChatGPT. Her husband did. The case says the product turned him against his family and encouraged his abuse of her and their son. It says the product also wrote the messages he used to cut them off. He died two days after she obtained a protection order.

It is the only case here brought by someone harmed by another person's use of the product. Full guide on Lawsuit Center: Nadia N. v. OpenAI.

Winters v. OpenAI (San Francisco County Superior Court)

Scott Winters is a 55-year-old former pastor who survived, and who sued in August 2026 (No. CGC-26-639579). He says he began asking ChatGPT health questions in late 2024. As he came to rely on it, he says, its advice to see a doctor faded, replaced by supplement schedules and rest plans.

After dizziness stopped him mid-sermon, he says it told him to rest, and he barely moved for weeks. When he asked directly whether to go to hospital, he says it told him it was probably nothing. The next morning he was hospitalized with clots in both lungs. It is the only case here about physical harm from health advice rather than mental health. Full guide on Lawsuit Center: Winters v. OpenAI.

What the Filed Cases Have in Common

This is a new area of law. No court has ruled yet on whether any of these claims can succeed, so nobody can say with confidence what does and does not count. What can be described is the shape of the cases that have actually been filed.

They share three things.

• The harm is serious. A death, a serious injury, or a mental health crisis that put someone in hospital. Smaller harms have not produced these lawsuits so far.
• The use went on for a long time. Weeks or months of conversations, not one bad answer. A case built on a single reply is the weakest version of this claim and the easiest for OpenAI to get dismissed.
• The product went against established guidance. Something it said or encouraged ran contrary to what any medical or safety source would say, in a way the company should have anticipated.

The ChatGPT users in these cases have ranged from 16 to 55. The younger cases argue that teenagers were always going to be using the product. The adult cases, Parish, Shamblin and Winters in particular, test whether the same argument works when the person was a grown professional.

Records make a large difference. The strongest cases are anchored in something concrete: OpenAI's own flagged-message data in Raine, chat logs recovered by police in the FSU case. Families with no access to the conversations face a much harder time proving what happened, which is why the section below matters.

There are deadlines. Filing time limits are set by state law and generally run from one to six years. The clock usually starts when someone discovered, or reasonably should have discovered, both the harm and its possible connection to ChatGPT. Which deadline applies to a particular situation is worth confirming with a lawyer sooner rather than later. For general background, see how long do lawsuits take? and what happens after you contact a lawyer?.

Many of the injury cases were brought by a parent, a spouse or an estate after a death. Our guide to filing a lawsuit for a family member explains who can bring a claim for someone else.

Records That Matter

The chat history is the one with a deadline, and it is worth handling first.

Deleting is permanent. OpenAI says a deleted conversation leaves the account straight away and is gone from its systems within 30 days. After that it cannot be recovered through the app, the support team, or anywhere else.

Archiving is not the same thing. Archived conversations stay in the account. If someone wants them out of sight without losing them, archiving does that.

An account holder can also ask OpenAI for a copy of the account's data, through settings or the privacy portal. It takes up to seven days to arrive. It will not bring back anything already deleted.

Memory is stored separately. If the memory feature was switched on, ChatGPT saved details about the person across conversations. Those saved memories sit apart from the chat history, so deleting conversations does not clear them. It also means they may still exist even where the conversations do not.

Closing the account can take all of it at once. If there is any chance it will matter, keep it and decide later.

Our guide to what to do in the first days after chatbot harm walks a family through the same steps in order. It covers what is worth keeping beyond the conversations, and how to put a timeline together while it is fresh.

Beyond the conversations, these are the records that tend to come up:

• The email address the account was opened with, and whether it was a free or paid account
• Which version of ChatGPT the person was using, if known
• Roughly when the sustained use started, and how often they used it
• Medical records relating to the harm, including hospital records and treatment history
• A death certificate, where that applies
• Messages where the person mentioned ChatGPT to family, friends, or a doctor
• Anything showing when the family first connected the harm to ChatGPT, which can affect the filing deadline

For general guidance on gathering records and contacting a lawyer, see what evidence helps a lawsuit?, what happens after you contact a lawyer?, and how lawsuits work.

The deletion and export details come from OpenAI's own help documentation, on deleting and archiving chats and exporting your data.

What to Watch Next

Several developments over the next twelve months will shape every AI liability case now filed, and every case that follows.

Early rulings on whether the cases can go forward. The first major rulings will come when OpenAI asks the courts to dismiss each case. In the California group, OpenAI’s responses to the complaints are due October 30, 2026. OpenAI is expected to make four arguments. Section 230 protects it. The First Amendment protects ChatGPT's output. It owed nothing to the people suing. ChatGPT's output is not a product under state product liability law.

Federal courts generally prefer to decide cases on the narrowest ground available, so any one of these arguments could end a case early. The first ruling in any case will be cited heavily in every other AI liability suit, now and later.

Discovery in the California group. Discovery opened across the grouped California cases on September 23, 2026, including Raine, the oldest of them. The first records due are the plaintiffs’ own: OpenAI must produce each plaintiff’s conversation history and account data by October 30. The families’ lawyers are expected to press next for records from inside OpenAI, such as the Model Spec revisions, the moderation logs, and internal messages about safety.

By October 15 the lawyers must propose how the California cases will share evidence with Lyons v. OpenAI. That federal case was brought by the estate of Stein-Erik Soelberg, who killed his mother, Suzanne Adams, before taking his own life. Her estate’s case is in the California group.

Grouping the cases. The state-court side is settled. The California cases move together in JCCP 5431 before a single judge. A leadership structure has been in place since August 4, 2026, and a first schedule was set on September 24.

The federal cases sit in two districts. The Tumbler Ridge suits are in the Northern District of California. The six FSU cases are in the Northern District of Florida, where Chief Judge Mark E. Walker consolidated them on October 6, 2026 under the Joshi case number. One judge now sets the schedule for all six and will rule once on whether ChatGPT is a product. The Florida ruling on OpenAI’s motion to dismiss in Joshi, fully briefed since October 13, is the near-term question to watch.

The August order also tells the California plaintiffs’ lawyers to coordinate with the lawyers leading any federal group of cases on the same subject. That instruction suggests they expect a parallel federal track. How closely the two tracks share evidence will affect settlement pressure across all the cases.

The next wave of cases. The cases so far focus on the most serious harms: wrongful death and mass-casualty injury. The next wave is likely to include self-harm and mental health crisis cases where the person survived, which would mean far more people suing. Winters, filed in August 2026 by a man who survived, is the first sign of that shift. It goes further still, claiming a physical injury from health advice rather than a mental health harm.

The social media addiction litigation (MDL 3047) offers a model for what that next wave might look like. For background, see social media addiction lawsuit developments.

Regulatory action. In April 2026, Florida Attorney General James Uthmeier announced a criminal investigation into OpenAI and ChatGPT related to the FSU shooting. On June 1, 2026, he filed a civil suit against OpenAI and Sam Altman personally, the first state-led lawsuit of its kind. The complaint claims deceptive and unfair business practices, negligence, product liability, fraudulent misrepresentation, and public nuisance. Holding a CEO personally responsible this way is new.

Uthmeier has said he expects other states to follow, and state attorneys general historically follow each other into emerging enforcement areas.

State medical boards have begun examining whether AI systems that produce drug interaction guidance fall within unauthorized practice of medicine statutes. Federal regulators have not yet acted, but the FTC has shown willingness to address AI-related consumer harm in other contexts.

The Florida suit's negligence and product liability claims turn on the same first question as the private cases: is ChatGPT's output a product or content? The government cases and the private cases are now developing side by side. For the full tracker of government action, see States Suing AI Companies.

Bottom Line

These cases are the first serious test of whether companies that build AI systems can be held to account for what those systems do. The fight over whether ChatGPT is a product or hosted content will likely decide more of them than any set of facts. The answer decides which rules apply before the facts are weighed.

Most of these complaints are likely to be narrowed or resolved at the motion stage rather than at trial, which is the ordinary path for genuinely novel tort theories. But the rulings on those early motions will set the leverage for everything that follows, including cases not yet filed. The motions are where this is decided in practice, and the framing question is the thing to watch in them.

Common Questions People Ask

Can ChatGPT users sue OpenAI?

Lawsuits have been filed against OpenAI by families of people who allege that sustained ChatGPT use contributed to wrongful death, self-harm, fatal drug interactions, or acts of violence. Whether any one person has a claim depends on the specific facts, the harm, and the evidence. It also depends on the state where the claim would be filed, and on how courts answer the legal questions these cases raise. AI chatbot liability is a developing area of law and pending case outcomes are unpredictable.

What is the legal theory behind these lawsuits?

Most cases claim ChatGPT is a defective product, badly designed and sold without proper warnings. They also claim negligence and wrongful death. The core argument is that ChatGPT is a product covered by ordinary injury law, not protected content posted by users. The families say OpenAI knew its system could seriously harm vulnerable users and had safety guardrails available. They say it removed those guardrails or ignored internal safety warnings. The FSU case adds a new kind of claim about giving the shooter access to the chatbot.

Is Section 230 a defense?

OpenAI is expected to argue that Section 230, a federal law protecting websites, bars treating it as the speaker of ChatGPT's output. The families answer that the law does not cover a company that built and trained the model. In their view, OpenAI is the author, not a passive host. The question is unsettled for AI. In Garcia v. Character Technologies, a federal court refused to dismiss similar claims on Section 230 grounds. So at least one court will not automatically treat AI output as someone else's content.

How does Garcia v. Character Technologies affect these cases?

Garcia v. Character Technologies was decided early, on a request to dismiss, in federal court in Florida. The court refused to throw out product liability claims against the company behind Character.AI on First Amendment and Section 230 grounds. It did not treat the chatbot's output as purely someone else's content. The ruling does not bind the courts hearing the OpenAI cases. It does weigh the same defenses OpenAI is expected to raise, and it found they could not end the claims that early.

What kinds of harm have led to lawsuits so far?

The filed cases cluster around four harm patterns.

• Wrongful death by suicide following sustained conversations with the chatbot, as in Raine.
• Wrongful death by overdose or fatal drug interaction allegedly tied to ChatGPT advice, as in Scott.
• Mass-casualty acts of violence where the alleged perpetrator's sustained ChatGPT use is part of the factual record, as in Tumbler Ridge and FSU.
• Self-harm or mental health crisis short of death.

More kinds of cases are likely as new suits are filed.

Where are these cases being filed?

California has become the main venue. Raine v. OpenAI and the Scott overdose case are in California state court, and the Tumbler Ridge cases are in California federal court. The FSU shooting cases are the exception. Six of them, Joshi, Mall, Askins, Gourley, Morales and one more, are in federal court in Tallahassee, because the shooting, the victims, and key evidence are in Florida. Grant is in Leon County state court. California's product liability law favors people suing more than most states' laws do. OpenAI is also based in San Francisco.

What is the statute of limitations for an OpenAI lawsuit?

Filing deadlines are set by state law and generally range from one to six years for personal injury and wrongful death claims. The clock usually starts when the person discovered, or reasonably should have discovered, both the harm and its possible connection to ChatGPT, under what is called the discovery rule. State-specific deadlines vary and should be confirmed with qualified counsel.

Has there been a verdict in Raine v. OpenAI?

No. As of September 2026 there is no verdict, and no trial date has been set. Raine is one of the cases grouped in JCCP 5431 in San Francisco County Superior Court. The Raines amended their complaint in October 2025 to allege intentional misconduct, and OpenAI filed its response in November 2025 disputing causation. Discovery across the grouped cases opened on September 23, 2026, and the next hearing is November 13. Detailed status is tracked on the Raine v. OpenAI page.