Listen to this article
Estimated 5 minutes
The audio version of this article is generated by AI-based technology. Mispronunciations can occur. We are working with our partners to continually review and improve the results.
Juno-award winning musician Ashley MacIsaac has ended his defamation suit against Google over its AI overview falsely labelling him as a sex offender.
Earlier this year, the 51-year-old fiddler had initiated a civil lawsuit with the Ontario Superior Court of Justice after he discovered the AI-generated summary at the top of Google’s search page — known as an AI overview — was asserting he had been convicted of a series of offences, including sexual assault, internet luring and attempting to assault a minor, and that he was on the national sex offender registry — all of which was untrue.
MacIsaac, who originally hails from Cape Breton, was seeking $1.5 million in damages from Google LLC.
“If a human spokesperson made these false allegations on Google’s behalf, a significant award of punitive damages would be warranted,” the lawsuit stated. “Google should not have lesser liability because the defamatory statements were published by software that Google created and controls.”
But on July 29, a notice of discontinuance was filed to end MacIsaac’s legal action. MacIsaac has not provided a reason for ending the lawsuit, but told the National Post he had no comment. CBC News has reached out for comment.
Google did not respond to CBC News’ request for comment on the discontinuance of the lawsuit.
Google Canada previously issued a statement in December saying its AI summaries shift frequently to provide the most “helpful” information, and that any mistakes are used to improve the system.
MacIsaac’s now-withdrawn claim lambasted Google for its “cavalier and indifferent response to its publication of utterly false statements,” stating that MacIsaac was owed a clear retraction.
Ashley MacIsaac performs during the East Coast Music Awards in Fredericton in 2022. The musician said the false claims caused him anxiety when performing. (Kelly Clark/The Canadian Press)
The company first began to introduce the AI overviews in 2024. They’ve since been the subject of controversy, sparking multiple lawsuits from companies and publishers who claim the feature has reduced traffic to their sites and criticisms from those who point out that AI is prone to hallucinating false information.
AI error led to cancelled concert
MacIsaac said in December that he found out about the misinformation — which he said came from online articles about a man in Atlantic Canada with the same last name — after a concert of his was cancelled by organizers who cited the AI summary.
MacIsaac later said the incident caused him to experience deep anxiety in subsequent performances, fearful that audiences may have seen the AI overview and his safety could be compromised.
In December, MacIsaac told CBC News that while he couldn’t afford to sue Google, many law firms had reached out to him offering to help out pro bono.
Justin Safayeni, a lawyer with Stockwoods LLP in Toronto who specializes in defamation law, told CBC News there are a number of reasons a plaintiff might discontinue a lawsuit before a case comes to trial: they may have run out of funds, believe the case to be unwinnable or not worth their time, or they may have chosen to settle privately with the other party.
Discontinuance is different from a dismissal in that it is possible to resurrect the case later, assuming there is no accompanying settlement that bars it, he said.
Courts grapple with question of responsibility
Google Canada previously issued a statement in December saying its AI summaries change frequently to provide the most ‘helpful’ information, and any mistakes are used to improve the system. (Annegret Hilse/Reuters)
The question of who exactly is responsible for damages to people, property or companies caused by false information hallucinated by AI is one that has dogged tech giants pushing the technology.
A Munich court issued a landmark judgement earlier this summer, finding that Google is legally liable for false information presented by its AI-generated summaries.
AI overviews were considered Google’s own content, the court ruled. The two German publishers behind the case said Google’s AI overviews had linked them to scams and promoted dubious business practices. Google said in June that it would challenge the ruling.
In 2024, Air Canada tried to argue its AI chatbot was a separate legal entity responsible for its own actions when faced with a lawsuit over its chatbot giving false information to a customer. The airline was ordered to take responsibility and pay compensation.
One of the reasons MacIsaac’s case has attracted so much attention is that Canadian courts haven’t tackled the question of liability when it comes to defamation resulting from AI statements yet, Safayeni says. But a reckoning is inevitable.
“Given the amount of use people are making of [AI tools], it’s going to come before the courts,” he said.
OpenAI, the creator of ChatGPT, admitted in a research paper last year that there is currently no way to stop AI large language models from hallucinating false information, because the models are built to generate only the most statistically plausible answers, with no discernment of what is actually true or false.