What Legal Risks Arise When AI Misrepresents a Company or Nonprofit?
When AI misrepresents a company or nonprofit, legal risk runs in two directions. First, the organisation may be harmed: an assistant falsely says it was sued, lost its licence or mistreated customers, and business or donations fall away. It may then have claims against the AI provider, or against the publisher of the source the answer repeated. Second, the organisation can create its own liability. Its own chatbot may give customers wrong information, or it may repeat false AI claims about someone else. The law on AI providers' liability is unsettled and moving quickly. For most organisations, the first steps are practical rather than legal: document, trace the source, correct it and report it. This page is general information, not legal advice.
Defamation vs misinformation vs opinion: what businesses need to know when AI search repeats a false claim
These three are often confused, and the difference decides what you can do:
| What it is | Example | Legal route | |
|---|---|---|---|
| Defamation | A false statement of fact about an identifiable organisation or person, communicated to others, that damages reputation | "The charity's director was convicted of fraud" when no such conviction exists | Possible claim, depending on jurisdiction, fault and damage |
| Misinformation | A false statement that is wrong but may not be defamatory, or not about you specifically | Wrong opening hours, an outdated price, a confused product feature | Correction requests; usually no legal claim |
| Opinion | A view that cannot be proved true or false | "Their prices are too high for what you get" | Generally protected, unless it implies false facts |
The rules differ by country. In the US, companies can bring defamation claims, but public figures, and some companies treated as such, must show a higher level of fault. In England and Wales, a body that trades for profit must show that the statement caused, or is likely to cause, serious financial loss. In the EU and UK, people named in answers, such as founders, directors and trustees, also have data protection rights, including the right to have inaccurate personal data corrected.
Can a business sue over a false Google AI Overview answer?
Possibly, but the law is still being tested, and recent cases point in different directions:
- Walters v. OpenAI (Georgia, 2025). A court granted OpenAI summary judgment in a defamation claim over a ChatGPT answer. The reasons included the warnings shown to users and the absence of proven damage.
- Wolf River Electric v. Google (Minnesota). A solar company sued Google after an AI Overview said it faced a lawsuit from the state attorney general, which it did not. A federal court sent the case back to Minnesota state court in January 2026.
- Starbuck v. Google (Delaware). A Delaware court refused to dismiss a defamation claim over statements generated by Google's AI, allowing the case to proceed.
- Munich Regional Court (Germany, May 2026). The court held that Google may be directly liable for false statements in AI Overviews, treating them as Google's own content rather than third-party material.
In the US, an open question is whether Section 230, which protects platforms from liability for third-party content, covers text that an AI system generates itself. Many commentators think it may not, but courts have not settled it.
Litigation is slow and expensive, and outcomes are uncertain. Most organisations get faster results by correcting sources and using the providers' reporting routes. Keep legal action for serious, persistent falsehoods that cause measurable harm.
What risks do organisations create for themselves?
- Your own chatbot. In Moffatt v. Air Canada (2024), a Canadian tribunal held the airline responsible for wrong refund information given by its website chatbot. If you deploy an assistant, you are generally responsible for what it tells customers.
- Repeating AI claims. Publishing or forwarding an AI answer that makes false claims about a competitor, a former employee or another organisation can expose you to the same claims as any other false statement.
- Marketing that relies on AI outputs. Claims generated by AI and published without checking, such as statistics, certifications or comparisons, carry advertising and consumer protection risk.
What specific risks do nonprofits face?
- Wrong status. Answers may say an organisation is not a registered charity or not tax-exempt. In the US, donors can check status through the IRS's Tax Exempt Organization Search. In England and Wales, they can use the Charity Commission's register. Point to these in corrections.
- Confusion with similarly named organisations, including ones involved in scandals or closures.
- Political attribution. Answers that wrongly attribute campaign positions to a US 501(c)(3) organisation can mislead donors and the public about an organisation that is barred from campaign intervention.
- Sensitive services. Misstatements about services for vulnerable people, such as eligibility, confidentiality or safeguarding, can cause real harm.
- Leaders' personal data. False statements about named staff or trustees engage their personal rights as well as the organisation's reputation.
What should you do first?
- Document the answer: the engine, date, exact question, full answer and cited sources, with screenshots. See how to document a false statement generated by an AI assistant.
- Trace the cause. Is the error in a cited source, or did the assistant combine sources wrongly? See how reputation teams can separate source errors from model synthesis errors.
- Correct the source, with evidence, if a publisher, directory or review site is wrong.
- Report to the provider. Use the in-product feedback and, for serious cases, the provider's legal or privacy request routes. Google, for example, has forms for reporting content for legal reasons.
- Publish an accurate statement on your own site, so answers that search the web have a clear source.
- Take legal advice if the statement is serious, persistent and causing measurable harm, and before sending formal letters.
What might an illustrative case look like?
Consider an illustrative food bank whose name is close to that of a charity closed after a financial scandal. Assistants asked about the food bank describe "an investigation into misused donations", citing news about the other charity. Donations through its website fall. The food bank documents the answers, confirms its registration and good standing on the charity register, and publishes a clear page distinguishing the two organisations. It asks two news sites to clarify their articles and reports the answers through each assistant's feedback and legal request routes. A solicitor advises that a claim would be slow and uncertain. Within two months, search-based answers stop repeating the confusion, and the food bank keeps the documentation in case it recurs.
What do people ask about AI misrepresentation and the law?
Is an AI provider liable for false answers?
It depends on the country and the facts, and courts are still deciding. A German court has treated AI Overviews as Google's own statements. US cases have gone both ways so far.
Do disclaimers protect AI providers?
They were one reason a Georgia court granted OpenAI summary judgment in Walters. Other courts may weigh them differently.
Should we send a legal letter straight away?
Usually not. Correcting sources and reporting through providers' channels is faster. Take legal advice before any formal letter.
What if the false claim is about our chief executive?
Treat it as both an organisational and a personal matter. The person may have data protection and defamation rights of their own.
Published by The Kovmere Team · updated Sep 29, 2026 · More resources