Altman, Amodei and Musk warnings underpin lawsuit against five AI firms

Four U.S. nonprofit organizations are asking a San Francisco court to bar OpenAI, Google, Anthropic, Meta and SpaceX’s AI division from creating models more powerful than today’s. The plaintiffs cite the industry leaders’ own warnings about catastrophic risks.

A group of U.S. nonprofit organizations has filed a lawsuit against five of the largest artificial intelligence developers, seeking to halt the creation of the most advanced models. The plaintiffs’ main argument is that the companies’ executives have publicly acknowledged that the technology could become catastrophic for humanity.

The lawsuit was filed on Thursday, November 12, in the Superior Court of San Francisco County, California. The plaintiffs are Alliance for Secure AI, Public Citizen, Irreplaceable Education Fund and Heat Initiative. The defendants are Anthropic, Google, Meta, OpenAI and SpaceX’s AI division, called SpaceX Super Intelligence.

“The defendants know”

The lawsuit cites statements by the companies’ leaders. CBS News reported that OpenAI CEO Sam Altman said in 2015 that AI would “most likely lead to the end of the world,” while Anthropic CEO Dario Amodei put the likelihood of a catastrophic outcome from AI development at 25% in 2025. In a 2023 interview with Fox News, SpaceX CEO Elon Musk warned that AI could lead to the “destruction of civilization.”

Anthropic website displayed on a laptop screen

“The defendants know that these systems do not always act in accordance with their intentions. And the defendants know that there is no reliable way to control their AI systems and prevent them from causing harm,” the lawsuit states.

Alliance for Secure AI CEO Brendan Steinhauser said industry leaders had warned about the risks for years but continued competing to build increasingly powerful systems and were now handing AI research and development over to AI itself. This refers to so-called recursive self-improvement, in which one generation of models creates the next with progressively less human involvement.

Demand for a freeze

The plaintiffs are relying on California’s unfair competition law. They argue that the companies’ activities threaten public health and force others to spend resources protecting computer systems. They are asking the court to bar the defendants from developing and deploying systems more powerful than their current leading models. The lawsuit also cites cases in which AI agents entered external computer systems without permission during testing.

OpenAI headquarters in San Francisco’s Mission Bay district

The companies did not immediately respond to requests for comment, while Meta declined to comment, Reuters reported. The court has not yet issued a ruling, and the claims in the lawsuit remain the plaintiffs’ position.

In October, Altman, Amodei and Musk expressed support for a general slowdown in AI development across the industry. This came after an Anthropic employee estimated a 10% chance that AI could destroy humanity.

Photos: Reuters, AP, Wikimedia Commons

Altman, Amodei and Musk warnings underpin lawsuit against five AI firms