Libel lawyer Paul Tweed says he is preparing a group legal action against technology providers including OpenAI, Meta, Google and Amazon, alleging that their AI chatbots and other AI content generating activities are breaching defamation and privacy laws.
Tweed, who has made his name as a defamation lawyer to the famous, including in May securing a €100,000 defamation payout from the BBC for former Sinn Fein president Gerry Adams, is building a group action, saying that content generated by the technology giants is new and doesn’t fall within the protections they have relied on to date.
Technology giants have been able to claim immunity from content produced on their social media platforms under s230 of the Communications Decency Act in the US, while in Europe, the EU’s Electronic Commerce Directive provides online intermediaries such as social media platforms a degree of immunity from liability for illegal content posted by users.
However, Tweed told Legal IT Insider that generative AI material produced by the likes of ChatGPT is new material that falls outside of this immunity. He is looking at three alleged grounds to bring an action: defamation by AI chatbots; unauthorised use of works for training AI models; and the creation by AI of fake biographies that he says are being sold by the likes of Amazon.
While the EU has strict laws around the use of AI, Tweed told us: “The UK and Irish governments have shown reluctance to bring in tight regulation, which is on one level understandable but I have people coming to me with serious allegations, and trying to get an AI platform to respond to anything take three weeks, by which time it’s too late.”
Tweed is trying to put together what he describes as a ‘Rubik’s cube’ claim that meets the necessary requirements in terms of jurisdiction, liability and funding. However, he says it’s “very likely” he’ll bring an action within the next six months, and at the moment he has a group of around 25 to 30 clients keen to move forward.
Meta, Google, Amazon and OpenAI have a presence in Ireland meaning, Tweed says, that he has jurisdiction and the companies are subject to European libel laws. However, he is considering all options as there are limited provisions for class actions in Ireland and they don’t allow third party funding.
In terms of the wider implications for people generating content from AI chatbots, Tweed warns that this is analogous to the early days of social media when people sharing libellous content didn’t fully appreciate that they would be liable for its content. “If you share libellous content you can’t say it was ChatGPT or AI’s fault,” he told us. “Once you do that, you’re a publisher.”
We have reached out to Meta, OpenAI, Google and Amazon for comment.
This potential action follows three separate cases in May in which lawyers included content from ChatGPT within their court submissions without checking its accuracy.
Three ChatGPT cases in May see US lawyers in hot water for error-filled briefs