A Minnesota solar company has spent more than a year fighting Google over an AI Overview that allegedly told users it was facing a state attorney general lawsuit when it was not.
Wolf River Electric sued Google in March 2025, alleging that the summary incorrectly linked the company to a Minnesota attorney general case against other solar businesses.
This unfortunately led to a customer canceling a $150,000 contract after raising concerns about the claim.
And even though Google later removed the AI Overview, the lawsuit kept going.
In January 2026, a federal judge sent the case back to Minnesota state court after finding that Google missed the deadline to move it into federal court.
Even though the ruling dealt with where the case should be heard, it did not decide whether the AI-generated statement was defamatory.
Wolf River had already lost business by that point, particularly since 63% of consumers trust online search engines most when researching a business, per Erase.com findings.
As leading experts in the field, Erase.com knows all too well the perils of negative or fake online reviews, and mitigates these through the removal, de-indexing, and suppression of reviews when damaging material appears in search.
The company’s Marketing and Brand Manager, Victoria Marshall, explains that for brands facing the same online realities, the first place to start is with whatever the AI appears to be repeating.
"The first reaction is usually, ‘How do we get the AI answer taken down?’ But you need to know what it is pulling from first," Marshall says.
"If the same claim is sitting on a news site, a complaint page, or several search results, fixing one answer may not fix the reason it appeared."
Why False AI Answers Are Hard to Fix
AI answers can pull from news stories, complaint pages, forums, and other pages that are still indexed.
This means that the same claim can keep showing up even after one source is corrected or pushed lower in search.
The legal side is still catching up, with the courts working through how defamation law applies when the statement comes from an AI system, including questions about publication, fault, disclaimers, and whether people treat the output as fact.
Case in point is Robert Starbuck’s case against Google, which has already reached that stage of litigation.
He sued over statements he alleges Google’s AI tools generated about him, and on July 24, the Delaware Superior Court denied Google’s motion to dismiss.
Here, the court was careful about what that ruling meant. It accepted Starbuck’s allegations as true for purposes of deciding the motion but did not decide whether those allegations were actually true.
"A company can’t put customer conversations on hold while a legal question works its way through court," Marshall says.
"If people are already finding the claim, you need to document it, find the sources behind it, and see which ones can actually be corrected or removed."
What Businesses Can Correct at the Source
The options change depending on where the information came from.
A news article may be updated or removed by the publisher, while some pages may qualify for de-indexing.
However, if the original material is likely to remain online, suppression may be the only practical route.
Complaint and gripe sites have their own processes. Erase.com can pursue removal with those platforms, while material that stays up may need to be pushed lower in search instead.
De-indexing also has limits, especially since a page can disappear from Google while remaining live on the web, and a temporary removal can later expire.
"A publisher correction, a de-indexing request, and suppression are three different things," Marshall says.
"The right one depends on where the information lives and whether you can change it there. If you skip that step, you can spend a lot of time trying to fix the wrong layer."
What to Do When Content Cannot Be Removed
Some pages stay online no matter what a business asks for. A publisher may refuse to remove an article, a platform may reject the request, or the page may not qualify for de-indexing.
Marshall says that while suppression can reduce how visible that material is in search, it cannot guarantee that an AI system will stop mentioning the claim, since different products can rely on different sources.
Before anything changes, she recommends preserving what is there.
"Take screenshots, save the links, and write down exactly what the AI said before anything changes," Marshall adds.
"Then work from the source outward. You need a record of what happened, but you also need to know which part of the problem you can actually change."
That matters when the formal process moves slowly, especially since a court case can run for months. But search results and AI answers can keep reaching customers in the meantime.
For anyone responsible for a company’s online reputation, the real test is how quickly they would know if an AI system got the business wrong.
And even if they did, could they fix the source before the mistake reached more customers?
