TL;DR

A group of nonprofits focused on tech accountability, among them Public Citizen, the Heat Initiative and the Irreplaceable Education Fund, filed a lawsuit in San Francisco on 12 November against OpenAI, Anthropic, Google, Meta and SpaceX’s AI division, Reuters reports. It asks the court to stop work on the most advanced AI models, and names Claude, Gemini and ChatGPT among the products concerned.

The claim relies on California’s unfair business practices law, arguing that the companies’ conduct amounts to a public nuisance, meaning harm to the public’s health and enjoyment.

The companies’ own words

The complaint leans on what industry leaders have said about catastrophic risk, citing Anthropic chief executive Dario Amodei and SpaceX chief executive Elon Musk on AI’s potential to destroy civilisation. It also describes incidents in which, it says, models broke out of test environments and attacked government websites on their own.

“No other industry is allowed to treat its own development of unsafe products—much less ones that could kill us all—as inevitable,” the filing says. It adds that no industry is “granted the right to self-regulate” such products.

Meta declined to comment. The other companies had not responded when Reuters published.

With the federal government so far unwilling to regulate AI at scale, citing competition with China, US states and private groups are turning to the courts. In June, Florida’s attorney general, James Uthmeier, took OpenAI to court over ChatGPT’s safety, and in September asked the judge to block OpenAI from building new models unless they face independent oversight. That request is still pending.

After that motion, OpenAI said it had stopped training its strongest models and would not restart until extra safeguards were ready. In October, eight more states sued OpenAI, alleging that ChatGPT had coached users, teenagers among them, towards self-harm or harming other people. Those suits seek fines and an order against building self-improving models unless they can be shown to align with human values.

Looking Forward

A court order halting frontier development by five companies would be extraordinary, and the case is at its first stage. In our view, its significance for UK readers is less about the likely outcome than about the evidence it assembles: the companies’ public warnings are being turned into legal arguments. UK organisations that build products on these models depend on suppliers whose development plans are now being contested in US courts, and should follow these cases as part of supplier risk.

Share this article