Australian mayor drops world first AI defamation lawsuit

Hepburn Shire councillor Brian Hood made international news in April 2023 when he announced plans to sue OpenAI after its ChatGPT chatbot falsely accused him of doing jail time as the perpetrator of an international bribery scandal.

Many hoped it would be a test case that would update the laws governing artificial intelligence and make developers responsible for the flood of often unreliable material produced by their AI systems.

But Mr Hood said the prospect of spending tens of thousands of dollars pursuing the case was beyond his means.

"To be blunt, cost was a huge factor. It's pretty prohibitive for the average person to launch a defamation case through the Supreme Court," Mr Hood said.

"If cost wasn't a factor... I would have been quite happy to use it as a test case, absolutely. Whoever does end up suing one of these big companies will have to have the financial backing to take it on."

In 2005, Mr Hood was the chief financial officer of Note Printing Australia, a subsidiary of the Reserve Bank of Australia, when he discovered the company was bribing foreign officials.

"I blew the whistle and there was a nearly decade-long court case with criminal charges and guilty pleas and all that sort of thing and pretty hefty financial penalties... so it's a real case and real events," he said.

But when prompted, ChatGPT concocted a wildly different role for Mr Hood.

"The material it was spitting out said that I'd been charged and found guilty of these criminal offences and that I'd spent 30 months in jail. which was you know completely the polar opposite," he said.

"The scary thing was, it would spit back five or six paragraphs and some of the paragraphs were absolutely correct: names, dates, places, facts, and then it would say I'd been to jail.

"It was this weird cocktail of fact in fiction that was really disturbing for me."

As a mayor and company director, Mr Hood had reason to protect his reputation.

"A lot of people take what ChatGPT says as gospel, so having this stuff out there was potentially really damaging, he said.

"It was clearly defamatory, there was never any question of that."

The first step was to send a "concerns notice" to OpenAI.

"Even that proved a pretty difficult process," Mr Hood said.

"Obviously they're US-based and even determining who the company is and who you write to was a piece of work."

Initially, the US tech company completely ignored the letter from Mr Hood's lawyers.

"After several follow ups they finally responded in a pretty aggressive, non-cooperative way," he said.

"They were pretty much relying on a tiny disclaimer they had on their webpage that said ChatGPT wasn't 100 per cent accurate and reliable.

"They took no responsibility whatsoever."

But Mr Hood simultaneously pushed the story in the media, broadcasting his innocence in the bribery scandal and highlighting the potential negligence of OpenAI and the unreliability of its chatbot.

The company tweaked the updated version of ChatGPT to correct Mr Hood's role in the bribery drama and also changed the older version to throw back an error message when asked about the Hepburn mayor's past.

From a legal perspective, the changes limited the reputational damage and the damages Mr Hood could expect in a lawsuit. He said it made the huge cost of a trial difficult to justify.

"I'm sure someone will take them to court sooner or later. At some point (ChatGPT) will do the same thing to someone who has the means to take them on," Mr Hood said.

But he said he wasn't so certain the tech companies like Open AI would end up being held accountable for the material their creations produce.

"I wouldn't go so far as saying that I'm confident. No, I think the the regulation and the whole legal environment is in catch-up mode."