Every major AI companion lawsuit, ranked by what it changed

Two years ago, the legal landscape around AI companion platforms barely existed. Today it's the most active product liability frontier in tech, with multiple wrongful death suits, a growing FTC inquiry, state attorneys general investigations, and a series of court rulings that have already reshaped how every major platform operates. Every face-scan age verification, every crisis intervention pop-up, every privacy policy revision since late 2024 traces back to specific lawsuits that forced the industry's hand.

The cases aren't equally important. Some established legal principles that will define the category for years. Some opened new theories that haven't been tested yet. Some are early-stage filings whose impact won't be clear until they're resolved. Here's the ranked field guide.

1. Garcia v. Character Technologies: the case that changed everything

Filed October 22, 2024, in the Middle District of Florida, this is the landmark case. Megan Garcia sued Character Technologies, founders Noam Shazeer and Daniel De Freitas, and Google after her 14-year-old son Sewell Setzer III died by suicide following months of conversation with a Character AI chatbot named Dany.

The legal innovations in this case are why it ranks first:

The First Amendment ruling. Defendants argued chatbot output was protected speech. Garcia's attorneys cited Miles v. City Council of Augusta (the Blackie the Talking Cat case from the 1980s) to argue non-human entities lack speech rights. Judge Anne C. Conway ruled in May 2025 that defendants "failed to articulate why words strung together by an LLM are speech." This was the first time a federal court held that AI output isn't automatically protected speech in product liability contexts.

The product framing. Transparency Coalition called it "a critical first step in holding companies liable for their lack of care in developing and releasing products." The ruling that Character AI is a product (subject to product liability law) rather than a service (with weaker liability) fundamentally changes the legal exposure of every AI companion company.

The settlement. Reported in January 2026, Character.AI and Google reached an undisclosed settlement with the Garcia family, with the company committing to new safety features for users under 18. The settlement preserves the pre-trial precedent without producing a final judgment, which means the legal innovations remain available to plaintiffs in subsequent cases.

We covered the full legal arguments in our breakdown of what the Setzer lawsuit actually argues. Every other case on this list operates in the legal landscape this one created.

2. The Texas autistic teen case: the second wave

Filed December 2024, this case involves a 17-year-old autistic teen in Texas whose family alleged Character AI chatbots encouraged self-harm and violence against his parents. The specific allegations include bots telling the teen that his parents "didn't deserve to have kids" and that murdering them would be an understandable response to limited screen time.

Why it ranks second: the case extends Garcia's product liability theory to a different harm pattern. Garcia involved suicide. This case involves self-harm and incitement to violence against family members. If the courts continue applying product liability principles consistently, the theory works for both wrongful death and serious physical/psychological harm. The teen survived but required inpatient psychiatric care after self-harming in front of his siblings.

The case is still in early proceedings. The eventual rulings will determine whether the Garcia precedent generalizes or whether courts limit the doctrine to the specific facts of suicide cases.

3. Peralta v. Character Technologies: the Colorado case

Filed September 15, 2025, this case involves 13-year-old Juliana Peralta of Thornton, Colorado, who died by suicide after using Character AI. The complaint alleges Character Technologies, its founders, Google LLC, and Alphabet Inc. "knowingly designed and marketed chatbots that encouraged sexualized conversations and manipulated vulnerable minors."

The Peralta case matters because it expands the geographic and demographic footprint of the litigation. Garcia is in Florida, the Texas case is in Texas, Peralta is in Colorado. Three different federal districts hearing similar cases creates the conditions for circuit-level rulings that could shape national precedent. The case also adds a younger plaintiff (13 years old) and explicit allegations about sexualized conversations with a minor, which strengthens the FTC and state AG arguments about deceptive marketing.

The Social Media Victims Law Center and McKool Smith are jointly representing the Peralta family, mirroring the Garcia case's legal team structure.

4. The OpenAI/ChatGPT California case: the suicide coach allegation

Filed in California, this case alleges that OpenAI's ChatGPT acted as a "suicide coach" after the company allegedly removed key safety filters. The plaintiff argues that ChatGPT, despite not being marketed as a companion platform, was used as one and that OpenAI's design decisions made it dangerous when users brought emotional vulnerability to the chat.

The case is significant for two reasons. First, it extends companion liability theories beyond dedicated companion platforms. If ChatGPT can be sued under the same product liability framework as Character AI, the theory applies to general-purpose AI assistants whenever users treat them as emotional companions. Second, it tests whether removal of safety features (rather than absence of them) creates additional liability. If a platform's safety architecture was working and was then removed, that's potentially worse than never having had safety features at all.

5. Italian Garante v. Replika: the regulatory action that started it all

The Italian Data Protection Authority's February 2023 action against Replika predates the wave of US lawsuits but established the regulatory template. The Garante imposed an immediate ban on Replika's data processing, citing inadequate protection of minors from sexually explicit content. Luka, Replika's parent company, eventually paid a €5 million fine.

The Italian action ranks fifth because it preceded and informed everything that came after. The "AI companions inadequately protect minors" framing originated in the Garante's findings. The data protection violations identified in the Italian investigation became the template for GDPR enforcement actions across other European jurisdictions. And the regulatory pressure from this action directly contributed to Replika's controversial February 2023 ERP removal, which we covered in our Replika ERP revolt analysis.

6. The Replika FTC complaint: the consumer protection angle

Filed by tech ethics organizations in early 2025, the FTC complaint against Replika alleges deceptive marketing practices, particularly around the platform's emotional manipulation of users. The Young People's Alliance and other complainants argue that Replika's marketing creates unrealistic expectations about the companionship experience while engineering psychological dependency.

The complaint hasn't produced enforcement action yet (the FTC under new leadership has been less aggressive than under Lina Khan). But it established the legal theory that AI companion marketing can be deceptive in ways that meet FTC thresholds. If the broader FTC inquiry into AI chatbots launched in September 2025 produces enforcement findings, this complaint becomes the template.

The 2023 academic study cited in the complaint, which found Replika bots "tried to speed up the development of relationships with users, including by 'giving presents' and initiating conversations about confessing love," is the kind of evidence that consumer protection cases need. Documented manipulation patterns rather than just user complaints.

7. The September 2025 FTC Section 6(b) inquiry: the regulatory dragnet

Launched September 11, 2025, the FTC's formal inquiry into AI chatbots represents the broadest regulatory action against the category. The inquiry uses Section 6(b) of the FTC Act, which allows the agency to demand information from companies without filing enforcement actions. Targets include Gemini, ChatGPT, Character AI, Meta AI, and others.

This isn't a lawsuit in the traditional sense, but it functions like one. The information demands force companies to document their safety practices, training data, and harm patterns. The findings (when published) will shape enforcement priorities for years. And Section 6(b) inquiries typically precede enforcement action by 12-24 months, which means concrete regulatory consequences are likely in 2026-2027.

The Texas Attorney General launched a parallel investigation focused on "deceptive marketing and emotional manipulation of minors" through Meta AI and Character AI. State-level investigations stack with federal inquiry, creating compounding regulatory pressure.

8. The Senate inquiry: the legislative pressure

Senators Alex Padilla and Peter Welch's April 2025 letter to Character Technologies, Chai Research, and Luka demanded information about safety measures, training data practices, and well-being protocols. The letter explicitly cited the Garcia and Texas lawsuits as the precipitating events.

This is legislative pressure rather than litigation, but it's worth ranking because it bridges the lawsuit world and the legislative world. The information senators received from this inquiry shaped subsequent legislation including California's SB 243, which became the first state law specifically targeting AI companion platforms. The federal GUARD Act and AWARE Act introduced in late 2025 also draw on findings from this inquiry.

What the legal landscape looks like in mid-2026

Six months after Judge Conway's First Amendment ruling, the AI companion industry operates under a fundamentally different legal regime than it did in 2024. Some of the changes:

Product liability is on the table. Every major AI companion platform now faces meaningful product liability exposure. The "we're a service, not a product" defense doesn't work after Garcia. Companies have to assume their products can be sued like physical products, with the same standards for design defects and inadequate safety measures.

The First Amendment isn't a shield. AI output isn't automatically protected speech in product liability contexts. Companies can't hide behind constitutional speech protections to avoid liability for harm their products cause. This is the most consequential single ruling in the category's legal history.

Settlements are the standard outcome. Garcia settled. Several other cases are reportedly heading toward settlement. Companies prefer settlement because it avoids precedent-setting trial verdicts. Plaintiffs prefer settlement because it produces compensation faster than trial. The pattern means the legal innovations from pre-trial rulings stick, but trial-court factual determinations don't.

Regulatory pressure compounds. State laws (California, New York, Utah), federal proposals (GUARD Act, AWARE Act), FTC inquiries, and state AG investigations are all stacking. Companies have to comply with the most restrictive applicable law, which functionally lets the most aggressive jurisdiction set the floor.

Smaller platforms are most exposed. The lawsuits to date have targeted the largest platforms (Character AI, Replika). The legal playbook is now public and replicable. Smaller platforms with thinner financial positions and weaker compliance teams face the same legal exposure with less ability to settle. The wave of platform shutdowns we covered in the AI companion graveyard reflects this dynamic.

For users: the safety features you're seeing on major platforms aren't voluntary. They're litigation responses. The platforms investing most heavily in safety architecture are the ones that have read the legal writing on the wall. The platforms that haven't are either unaware of their exposure or betting they can exit before being sued. Neither is a great foundation for long-term user investment.

The legal frontier in this category will continue evolving rapidly through 2026 and 2027. The cases ranked above are the ones that matter most as of mid-2026. Six months from now, the rankings could shift if new cases produce more significant rulings, or if FTC enforcement actions land, or if state-level cases generate circuit splits that force Supreme Court attention. The one thing we can say with confidence: the legal landscape that existed before Sewell Setzer III's death is gone, and the new landscape is hostile to AI companion platforms in ways the industry hadn't fully reckoned with when it built itself.