An Australian mayor said he may take legal action over false information shared by advanced chatbot ChatGPT.
Brian Hood, Mayor of Hepburn Shire Council, says the OpenAI-owned tool falsely claimed he was imprisoned for bribery while working for a subsidiary of Australia's national bank.
In fact, Mr Hood was a whistleblower and was never charged with a crime.
His lawyers have sent a concerns notice to OpenAI - the first formal step in defamation action in Australia.
OpenAI has 28 days to respond to the concerns notice, after which time Mr Hood would be able to take the company to court under Australian law.
If he pursues the legal claim, it would be the first time OpenAI has publicly faced a defamation suit over the content created by ChatGPT.
OpenAI has not responded to a BBC request for comment.
Millions of people have used ChatGPT since it launched in November 2022.
It can answer questions using natural, human-like language and it can also mimic other writing styles, using the internet as it was in 2021 as its database.
Microsoft has spent billions of dollars on it and it was added to Bing in February 2023.
'Plausible-sounding but incorrect'
When people use ChatGPT, they are shown a disclaimer warning that the content it generates may contain "inaccurate information about people, places, or facts".
And on its public blog about the tool, OpenAI says a limitation is that it "sometimes writes plausible-sounding but incorrect or nonsensical answers".
In 2005, Mr Hood was company secretary of Notes Printing Australia, a subsidiary of the Reserve Bank of Australia.
He told journalists and officials about bribery taking place at the organisation linked to a business called Securency, which was part-owned by the bank.
Securency was raided by police in 2010, ultimately leading to arrests and prison sentences worldwide.
Mr Hood was not one of those arrested, and said he was "horrified" to see what ChatGPT was telling people.
"I was stunned at first that it was so incorrect," he told Australian broadcaster ABC News.
"It's one thing to get something a little bit wrong, it's entirely something else to be accusing someone of being a criminal and having served jail time when the truth is the exact opposite.
"I think this is a pretty stark wake-up call. The system is portrayed as being credible and informative and authoritative, and it's obviously not."
Different chatbots, different answers
The BBC was able to confirm Mr Hood's claims by asking the publicly available version of ChatGPT on OpenAI's website about the role he had in the Securency scandal.
It responded with a description of the case, then inaccurately stated that he "pleaded guilty to one count of bribery in 2012 and was sentenced to four years in prison".
But the same result does not appear in the newer version of ChatGPT which is integrated into Microsoft's Bing search engine.
It correctly identifies him as a whistleblower, and specifically says he "was not involved in the payment of bribes... as claimed by an AI chatbot called ChatGPT".
