The New Bill and Its Penalties
Sen. Bernie Sanders and Rep. Greg Casar introduced a bill to ban artificial “superintelligence.” The 19-page Ban Artificial Superintelligence Act would outlaw recursive self-improvement AI. This technology could allow machines to train themselves to get smarter without human control. Violating these new rules carries a corporate death penalty and up to 20 years in jail. Sanders said this penalty matches the punishment for unlawfully building nuclear weapons.
The bill creates a Cabinet-level Department of Artificial Intelligence to regulate the tech. Companies must pause development until this new agency is fully established. NBC News reported that the bill defines superintelligence as the capacity to automate or greatly accelerate AI research. The goal is to stop systems that might eventually escape human control. New York Post noted that some AI specialists warn of these exact risks.
Sanders’ Warning on Safety
On Wednesday, Sanders informed reporters via a call that virtually nothing is happening to prepare us for the tsunami that is coming. He argued that the current lack of safety guardrails poses dangerous risks. “We’re dealing with something that humanity has never seen before,” Sanders said. He called it the most transformational technology in history moving very rapidly. The bill aims to set a global standard for how this tech is used.
While this bill moves forward, other Democrats are pushing for talks with China. A letter from 17 senators led by Sen. Chris Van Hollen urges cooperation on safety. They want a formal agreement to slow down frontier AI development. Washington Times reported that two leading left-wing lawmakers are proposing this federal department. This year, however, meaningful legislation is not expected to pass.
The proposed agency would be led by a secretary of AI. This official would supervise the pause and the new safety rules. NBC News stated that the bill requires companies to stop work until the agency is created. The penalty for breaking these rules is severe. It is designed to stop the race between the U.S. and China. Sanders maintains that global leadership in AI safety is essential.
The bill lists specific precursors to define superintelligence development. These include the ability to speed up the research process significantly. New York Post confirmed the bill targets recursive self-improvement models. The aim is to prevent AI from learning faster than humans can manage. Washington Times noted the focus on models that escape human controls.
