ai chatbot lawsuit: what the adjudicated record shows in 2026

Three courts on two continents have now answered the same question: when a chatbot gets it wrong, who pays? The consistent finding is the operator, not the bot.

Anyone who has put a chatbot on a customer-facing site is now carrying a question they may not have priced: if it says something wrong, who is on the hook? The vendor pitch implies the model is a tool you point at your content. The case law says something else.

It says the chatbot is you. When Air Canada’s website bot invented a refund policy, the airline argued in a tribunal that it “is a separate legal entity that is responsible for its own actions.” The member’s reply was two words: “remarkable submission” (source). Air Canada was ordered to pay C$812.02 (source).

That was 2024, and it was not a one-off. Since then a German court has ordered a clinic to stop after its chatbot fabricated medical credentials, and another German court has held Google directly liable for a defamatory AI search summary. Read together, the rulings converge on one point, and it is not the point most headlines take.

What an ai chatbot lawsuit is

An ai chatbot lawsuit is a legal claim that a deployed conversational AI caused harm, through wrong advice, a false statement, or a defamatory answer, and that the company that operated it is legally responsible for the output. The disputes turn on a single question the courts keep answering the same way: is the chatbot’s answer the operator’s own act, or a third party’s?

Can a company be held liable for what its chatbot says?

Yes, and the defence that it cannot has failed on both sides of the Atlantic. Two courts have now rejected near-identical versions of the argument that an autonomous bot is a separate actor the operator cannot be blamed for.

In Moffatt v. Air Canada, the BC Civil Resolution Tribunal held that “It should be obvious to Air Canada that it is responsible for all the information on its website. It makes no difference whether the information comes from a static page or a chatbot” (source). The airline was found liable for negligent misrepresentation because it “did not take reasonable care to ensure its chatbot was accurate” (source).

Two years later a German court reached the same result under different law. On 12 May 2026 the Higher Regional Court of Hamm (Az. 4 UKl 3/25) held that a chatbot’s answers were the operating company’s own misleading commercial acts, not the conduct of an independent third party (source). As the reporting legal press put it, “the chatbot is not a third party within the meaning of the law” (source).

What was the Air Canada chatbot lawsuit?

It was the first widely reported case to test who owns a chatbot’s answer, and the customer won. In November 2022 Jake Moffatt used Air Canada’s support chatbot while booking flights after his grandmother died. The bot told him bereavement rates could be claimed retroactively, which directly contradicted the airline’s own bereavement-travel page that the bot itself linked (source).

The tribunal ordered C$812.02: C$650.88 in damages, C$36.14 in pre-judgment interest, and C$125 in fees (source). The dollar figure is small. The principle it set, that a company cannot outsource accountability to its own bot, is what makes the case cited far out of proportion to its damages.

Why did the German clinic case matter more than its size suggests?

Because the court explained the mechanism of liability, and it was control, not intent. The clinic’s chatbot told visitors its two managing directors were “specialists in plastic and aesthetic surgery”, a qualification they do not hold, and two of the titles it used do not exist as recognised German qualifications at all (source).

The company argued the bot ran autonomously on statistical probabilities and was not steered answer by answer, so its outputs were not attributable to the company (source). The court found the opposite: the decisive factor was that the company set the bot’s operating framework and had enough influence over it, shown by how easily it later reprogrammed the bot to stop making the false claims (source). It ordered a cease and desist under threat of an administrative fine of up to 250,000 euros for each violation (source). The ruling is not final: because attributing AI statements raises new legal questions, the court admitted an appeal to Germany’s Federal Court of Justice (source).

Are AI search summaries covered too?

In at least one German ruling, yes, and the reasoning is the same attribution logic applied to a search engine. On 28 May 2026 the Landgericht Muenchen I held Google directly liable for a defamatory AI Overview that linked two Munich publishers to a fraud scheme (source).

The court ruled that an AI-generated overview “is not a mere display of search results, but rather content attributable to” the search operator (source). That removed the liability shield search engines normally rely on for third-party results. The court issued a preliminary injunction enforceable by a fine of up to 250,000 euros per violation (source). It is a preliminary-injunction proceeding, not a final judgment on the merits, and can be appealed (source).

What about the chatbot cases that never reached a ruling?

Some of the loudest pressure on chatbot makers has not produced a verified outcome, and this post will not pretend otherwise. Character.AI removed open-ended chat for users under 18 on 29 October 2025, a retreat from its core companion product, taking effect no later than 25 November 2025 (source). TechCrunch reported the company was “facing lawsuits and public outcry after at least two teenagers died by suicide following prolonged conversations with AI chatbots” on the platform (source).

That is a product walk-back under litigation pressure, not a court holding. TIN’s record verifies the reversal and the reported context, not the outcome of any wrongful-death suit, which remains live. It sits here as the boundary of what the adjudicated record can say: a company changed its product, and the suits that pushed it have not yet been decided.

The four cases side by side

Case	Court and country	What the chatbot did	Who was held responsible	Penalty or status
Moffatt v. Air Canada	BC Civil Resolution Tribunal (Canada)	Invented a retroactive bereavement-fare policy	The airline, not the bot	C$812.02 ordered; airline complied
Verbraucherzentrale NRW v. Aesthetify	Higher Regional Court of Hamm (Germany)	Called directors “specialists” they were not	The operating company	Cease and desist, up to 250,000 euros per violation; appeal to Federal Court of Justice admitted, not final
Publishers v. Google	Landgericht Muenchen I (Germany)	AI Overview called publishers a fraud scheme	Google, as its own content	Preliminary injunction, up to 250,000 euros per violation; not a final judgment
Character.AI (under-18 chat)	No court ruling	Companion chat linked to teen-safety suits	Not adjudicated	Company removed under-18 open-ended chat; suits still live

What TIN’s case files document

Each row above is a TIN case file built on the primary record, not a press summary. The Air Canada figures are quoted from the signed tribunal decision, corroborated by two independent newsrooms. The Hamm holding is taken verbatim from the OLG Hamm judgment, with the appeal to the Federal Court of Justice noted. The Munich ruling is drawn from the official Bavarian court document, flagged as preliminary. The Character.AI reversal is bound to the company’s own announcement and two newsrooms, and is honest that the underlying suits are unresolved.

That is the difference between this page and a listicle of chatbot lawsuits. Every figure here can be followed back to the document it came from, and every case that is not settled says so.

The bottom line

The tempting read is that courts are turning against AI chatbots. The record does not support it. Two of these bots were ordered stopped, one search feature was enjoined, and one product was pulled, but none of it was a ruling against the technology. Each turned on a narrower and more durable question: whose speech is a machine’s output?

Every court that has answered gave the same answer. The output belongs to whoever deployed it. The “autonomous system, not our act” defence was raised in Canada and in Germany and lost both times, and a search engine’s usual shield for third-party content did not survive contact with a summary the engine wrote itself.

For an operator, the transferable lesson has nothing to do with AI law. If you put words in front of a customer, they are your words, whatever generated them. The exposure is not the model, it is the absence of a check between the model and the customer, and that check is a thing you can build before a tribunal builds it for you.

Sources

1. BC Civil Resolution Tribunal, Moffatt v. Air Canada, 2024 BCCRT 149, 2024-02-14. https://www.canlii.org/en/bc/bccrt/doc/2024/2024bccrt149/2024bccrt149.html
2. Ars Technica, “Air Canada must honor refund policy invented by airline’s chatbot,” 2024-02-16. https://arstechnica.com/tech-policy/2024/02/air-canada-must-honor-refund-policy-invented-by-airlines-chatbot/
3. Oberlandesgericht Hamm (Justiz NRW / NRWE), 4 UKl 3/25, full-text judgment, 2026-05-12. https://nrwe.justiz.nrw.de/olgs/hamm/j2026/4_UKl_3_25_Urteil_20260512.html
4. Legal Tribune Online (dpa), “OLG zur Haftung fuer irrefuehrende KI-Chatbot-Aussagen,” 2026-05-12. https://www.lto.de/recht/nachrichten/n/olg-hamm-aerzte-haften-chatbot-falsche-facharztbezeichnung-ki-uwg
5. gesetze-bayern.de (Bayerische Staatskanzlei), LG Muenchen I, 26 O 869/26 (BeckRS 2026, N 11860), 2026-05-28. https://www.gesetze-bayern.de/Content/Document/Y-300-Z-BECKRS-B-2026-N-11860?hl=true
6. PPC Land, “Munich court holds Google liable for AI Overviews defamation, a first,” 2026-06-24. https://ppc.land/munich-court-holds-google-liable-for-ai-overviews-defamation-a-first/
7. TechCrunch, “Character.AI is killing the chatbot experience for minors,” 2025-10-29. https://techcrunch.com/2025/10/29/character-ai-is-killing-the-chatbot-experience-for-minors/
8. TIN case files: Moffatt v. Air Canada, OLG Hamm 4 UKl 3/25, LG Muenchen I 26 O 869/26, Character.AI under-18 chat removal (full source lists on each page).

Questions

Can a company be sued for what its AI chatbot says?

Yes, and companies have lost. A BC tribunal ordered Air Canada to pay C$812.02 after its website chatbot gave wrong bereavement-fare advice, and a German court ordered a clinic to stop after its chatbot invented specialist medical titles. Both courts rejected the argument that the chatbot was a separate party the operator was not answerable for.

Who is liable when an AI chatbot gives wrong information?

The operator is liable when its AI chatbot gives wrong information. In every adjudicated case on TIN's record the court attributed the chatbot's output to the company that deployed it, not to the software, the vendor, or the model. The 'the AI did it' defence has failed each time it was tried.

What was the Air Canada chatbot lawsuit?

The Air Canada chatbot lawsuit was Moffatt v. Air Canada, decided by the BC Civil Resolution Tribunal in February 2024. The airline's website chatbot told a passenger he could claim a bereavement fare retroactively, which contradicted the airline's own policy page. The tribunal found negligent misrepresentation and ordered C$812.02.

Are AI search summaries treated the same as chatbots in court?

In at least one German ruling, yes. The Landgericht Muenchen I held that a Google AI Overview is Google's own attributable content, not a mere display of search results, and issued an injunction over a defamatory summary. That is a preliminary injunction, not a final judgment.

Sources

1. BC Civil Resolution Tribunal (CanLII), Moffatt v. Air Canada, 2024 BCCRT 149 , 2024-02-14
2. Ars Technica, Air Canada must honor refund policy invented by airline's chatbot , 2024-02-16
3. Justiz NRW / NRWE, Oberlandesgericht Hamm, 4 UKl 3/25 (full-text judgment of 12 May 2026) , 2026-05-12
4. Legal Tribune Online (dpa), OLG zur Haftung fuer irrefuehrende KI-Chatbot-Aussagen , 2026-05-12
5. gesetze-bayern.de (Bayerische Staatskanzlei), LG Muenchen I, decision in case 26 O 869/26 (BeckRS 2026, N 11860) , 2026-05-28
6. PPC Land, Munich court holds Google liable for AI Overviews defamation, a first , 2026-06-24
7. TechCrunch, Character.AI is killing the chatbot experience for minors , 2025-10-29

---

This is analysis, not a verified outcome. It carries no verification badge and never will. The proof lives in the case files, where every figure is checked against the public record and the method is printed on the page.