A regional mayor may take unprecedented legal action to sue the creator of AI chatbot ChatGPT for defamation, after the program falsely claimed he was a guilty party in a foreign bribery scandal.
Brian Hood, the mayor of Hepburn Shire Council, west of Melbourne, was recently alerted to the fact that ChatGPT, a chatbot designed to provide detailed responses to users' questions, had incorrectly described his role in a foreign bribery incident in the early 2000s.
Councillor Hood did previously work for the company involved, Note Printing Australia, but was actually a whistleblower who told authorities about bribe payments being made to foreign officials to win contracts.
Media law specialist Professor David Rolph said suing an "online intermediary" for defamation would be complicated, as matters of jurisdiction would need to be taken into account.
If Cr Hood proceeds with legal action, it is believed it will be a landmark case in Australia to determine if artificial intelligence companies can be held liable for false information dispensed by their chatbots.
Cr Hood told the ABC today that he not only exposed Note Printing Australia's actions, he "became a prosecution witness and went through all of that process through numerous court cases".
"According to ChatGPT, I was one of the offenders, that I got charged with all sorts of serious criminal offences. I was never charged with anything," he said.
When a user asked ChatGPT about Cr Hood or his role in the bribery incident, it would incorrectly describe him as a guilty party in the scandal instead of saying he was the whistleblower.
Cr Hood said he had tendered a "concerns letter" to OpenAI, the company that created ChatGPT.
Under Australian law, a concerns notice must be issued by an "aggrieved person" explaining the situation and requesting action to be taken to rectify the issue in some way within 28 days.
"We haven't had any responses yet … I was shocked, I was really taken aback. I had to read it and read it again," he said.
"What was really disturbing, was that some of the paragraphs were absolutely correct and precise. They had the right facts, figures, names, dates, places, and all that.
"It was always very black and white as to what my role was. And I gave evidence in a whole number of court cases and a parliamentary inquiry."
OpenAI has been contacted for comment.
Legal 'burdens' in suing chatbots
Professor Rolph, a media law specialist at the University of Sydney, said Mr Hood could face a few "different issues" in any legal proceedings.
"One of the issues that we have with a lot of online intermediaries is the basic question of jurisdiction … can you actually bring a proceeding against them in an Australian court?" Professor Rolph explained.
"A lot of these internet intermediaries are based offshore, a lot of them in the United States, which will often raise all sorts of problems."
Professor Rolph said the legalities around AI technologies, such as ChatGPT, were still uncharted.
"Those sorts of technologies do pose a lot more burdens on people who want to sue for defamation," he said.
"And I think that's a function of the nature of the technologies.
"Now, I think it's much more difficult because these technologies are so dynamic, and so sort of variable … And they create more forensic burdens on people who want to sue to protect their reputations."
OpenAI's website states that "our text models are advanced language processing tools that can generate, classify, and summarise text with high levels of coherence and accuracy".
An academic paper written by several researchers from Georgetown University's Center for Security and Emergency Technology, examines the potential implications of language models such as ChatGPT in creating "misleading text".
The paper recommends a more "cooperative approach" between AI developers, social media companies and government agencies could help chatbots to avoid releasing misleading information.
