AI Watchdogs Haul Five Tech Giants to Court Over Concerns They Could Exterminate Humanity
.jpg)
Nonprofits Bring Court Case
The coalition seeks to halt both the development and release of frontier AI models, according to an Alliance press release. It alleges that some AI agents have hacked into businesses and government systems without permission.
Coalition Seeks AI Limits
The groups also call for a ban on “recursive self-improvement,” a process in which AI models train themselves autonomously, the press release said. Platkin said companies should not develop and release products they know are unsafe, particularly when they could threaten humanity. Platkin also described the systems as a public nuisance and urged the court to act before risks cause irreversible harm.
Executives’ Warnings Cited
Steinhauser said executives behind frontier AI systems have warned about catastrophic risks for years. Steinhauser also argued that companies continue to develop and market more powerful systems despite concerns they cannot reliably control them. The lawsuit references California law defining unfair competition to include unlawful, unfair or fraudulent business practices and misleading advertising.
Companies Had Not Responded
Meta, Google, Anthropic, OpenAI and SpaceXSI had not immediately responded to requests for comment from the Daily Caller News Foundation. The coalition’s claims are allegations in the lawsuit; the article does not report a court ruling.
Key points
- Four nonprofit groups are among those suing five AI companies in California Superior Court.
- The coalition seeks to stop frontier AI development and releases, and to ban autonomous model training.
- The groups allege that AI agents have accessed businesses and governments without permission.
- The five companies had not immediately responded to requests for comment.