Sanders and Casar propose banning artificial superintelligence with prison penalties

Under the proposed law, AI leaders who continue to develop artificial superintelligence despite the ban could face up to 20 years in prison. This penalty is intended to be comparable to existing penalties for the unlawful development of nuclear weapons and would also apply to "rogue actors" operating outside of AI companies.

The legislation also calls for a pause on the development of "frontier models"—advanced AI systems trained on specific thresholds of data—until a scientist-led Department of Artificial Intelligence is created. This new agency would be responsible for monitoring frontier models, overseeing the removal of dangerous features, and supervising the "destruction" of any artificial superintelligence.

Under the bill, AI companies would also be required to obtain permission from the Department of AI before developing any advanced AI models. Sanders stated in a press release that the legislation is necessary to prevent Big Tech CEOs from writing their own rules when the future of humanity is at stake.

PLUS ULTRAby Amenoyomi

The bill's severe penalties are rooted in the classification of artificial superintelligence as an existential threat. By defining this technology as capable of overthrowing the government or causing the destruction of humanity, the legislation treats its unauthorized development with the same legal gravity as the unlawful creation of nuclear weapons, prescribing up to 20 years in prison for violators.

To prevent such risks, the proposal establishes a tiered intervention strategy. It calls for an immediate pause on the development of "frontier models"—advanced systems defined by a specific threshold of training data—effectively halting the path toward superintelligence. Development of these advanced systems would only resume under a permission-based system managed by a newly created, scientist-led Department of Artificial Intelligence.