Nonprofits sue five AI companies to stop frontier models
The groups want a San Francisco court to stop new AI models until the companies show control, not damages.
On Thursday, four nonprofit groups with the Alliance for Secure AI in them sued five large AI companies in San Francisco Superior Court. They want the court to stop work on frontier models, which are the new large AI systems. The groups tell that the companies know they cannot control these systems, and the suit wants no damages.
What the suit says
The suit names these companies:
The groups name recursive self-improvement: AI models that make and train the next AI systems, with small control by a person. They tell that this makes the time short between a fault and its effect at large scale.
The numbers
The suit uses the words of the leaders of the companies. In 2015, Sam Altman told that AI “would probably most likely lead to the end of the world”. In 2025, Dario Amodei gave a 25 percent risk of very bad results. The suit also tells of AI agents that did work that no person approved.
What is not known
The source gives no words from the companies and no date for a court order. Anthropic will set its share price on 18 November. The groups want to stop new work before the end of the year.
Insight: The court action Anthropic wanted now points at it
In October, Anthropic wanted the California attorney general to go to court to stop a model that can cause large damage. At this time, the same court step points at all five companies, and the suit uses the words of their leaders about the risk. This seems to show that risk words from a company can be used against it. Vendor contracts must say who can stop the AI.
Sources
Posted
