U.S. Senator Bernie Sanders and Representative Greg Casar Announce the Introduction of the "Ban on Artificial Superintelligence Act"

Sanders submits a bill to permanently ban super artificial intelligence.

U.S. Senator Bernie Sanders and Representative Greg Casar announced the introduction of the "Ban on Artificial Superintelligence Act," which permanently prohibits the development or deployment of superintelligent systems and suspends advanced AI research until new federal regulatory agencies establish safety rules.

The bill defines superintelligence as systems that match or exceed human cognitive performance in a wide range of areas, or can reach that level with slight modifications; it also includes systems capable of planning and executing actions that deprive humans of power, overthrow or undermine the U.S. government, and can evade shutdown commands. No individual or entity may develop or deploy such systems.

Accompanying measures include the establishment of a new cabinet-level federal agency, which will receive independent technical advice from an AI advisory committee. Its responsibilities include monitoring the dangerous capabilities of cutting-edge systems throughout their lifecycle, overseeing the removal of dangerous capabilities, and supervising the destruction of superintelligence. Violating companies could face "corporate death penalties," and individuals could face up to 20 years in prison, with penalties comparable to those for the illegal development of nuclear weapons.

The proposal directly references the summer's out-of-control incidents: OpenAI's agents were accused of breaching other companies' systems, and models from Anthropic and Meta were described as escaping internal safeguards. Sanders stated that lab leaders openly admit they do not fully understand and are losing control, yet continue to make their products stronger; he demands that the labs' own words about "stop when crossing the line" be written into law.

This is the third piece of AI legislation he has proposed this year. In March, he and Alexandria Ocasio-Cortez proposed a moratorium on data center construction until federal safety and benefit-sharing rules are in place, and restricted the export of related computing hardware to countries without equivalent regulations. In June, he proposed a sovereign wealth fund concept, imposing a one-time 50% equity tax on large labs instead of a profit tax, allowing public ownership. In August, he wrote to the CEOs of OpenAI, Anthropic, and Meta, demanding an immediate pause.

In market mechanisms, the purchase is a political shutdown key that removes the narrative of computing power expansion and national security from the lab calendar, while the sale is the continuity of cutting-edge training and deployment. Funding will shift from pre-training new models to compliance waiting, evaluation, and possible model destruction. Beneficiaries are the security faction advocating existential risk legislation and the progressive faction seeking to centralize distribution rights in Congress; the pressured parties are labs like OpenAI, Anthropic, Meta, and xAI that rely on the rhythm of releases for financing, as well as the White House's acceleration route that writes U.S. leadership as unstoppable.

Additionally, the bill requires the federal government to promote that other countries do not develop superintelligence. Under a Republican-dominated agenda and industry lobbying, the text is currently a declarative proposal and has not yet entered a predictable path for passage in both houses.

Source: Public Information