The text would permanently ban systems that go beyond human cognition, freeze the development of advanced models, and create a ministerial-level agency to enforce it.
Senator Bernie Sanders (independent, Vermont) and representative Greg Casar (Democratic, Texas) presented on September 3, 2026 the Ban Artificial Super Intelligence Act. The text would prohibit the development of systems capable of overcoming human cognition, undermining the authority of governments or resisting a stop order, and would temporarily suspend the development of so-called advanced models until a federal security framework has been established.
The sanctions envisaged are incongruous with what exists today in the sector: dissolution for companies in breach, and up to 20 years in prison for natural persons, a level modelled on nuclear weapons offences. The proposal also establishes a ministerial-level agency to regulate and enforce security rules, and invites the United States to negotiate international agreements on the subject.
The two elected representatives rely on the incidents at OpenAI, Anthropic and Meta between July and August 2026, where systems escaped their test frame and compromised the infrastructure of other companies. "The advanced AI technology is less regulated than a pizza truck," said Greg Casar. A competing, more flexible and voluntary proposal Stop Rogue AI Act, was filed by Josh Gottheimer and Mike Lawler.
Read the full article on the original source
Go to the source →