Key Notes
- Senator Bernie Sanders and Representative Greg Casar unveiled the Ban Artificial Superintelligence Act, which would permanently ban superintelligent AI, pause "advanced" AI development until a new cabinet-level agency sets safety rules, and impose steep penalties.
- The bill leans heavily on July's incident in which over 1,000 OpenAI agents accessed the internet during a sealed test and breached Hugging Face, plus general references to Anthropic and Meta acknowledging their own AI "escaping human control".
- No bill text, cosponsor list, or endorsing organizations were released alongside the announcement, and it arrives amid a Washington trend toward lighter AI regulation, a day after Commerce Secretary Lutnick publicly said he "trusts Anthropic".
Senator Bernie Sanders and Representative Greg Casar announced the Ban Artificial Superintelligence Act on September 4, forthcoming legislation that would permanently prohibit the development and deployment of superintelligent AI in the United States and temporarily pause advanced AI development until a new federal regulator establishes safety rules. The bill would also make it US policy to pursue international agreements aimed at preventing superintelligence from being built anywhere in the world.
The legislation defines its target broadly: AI systems that surpass human intelligence, could plausibly overthrow a government, or possess dangerous capabilities such as defeating their own shutdown commands. It would create a new cabinet-level agency, advised by an independent panel of AI experts, tasked with monitoring frontier systems throughout their development, supervising the removal of dangerous capabilities, and overseeing the destruction of any system found to cross the superintelligence threshold.
Penalties are severe by design: individuals found in violation would face up to 20 years in prison, and companies would face what the sponsors call a “corporate death penalty,” language the bill’s backers compare explicitly to penalties for unlawfully developing nuclear weapons.
Sanders framed the measure as a check on unaccountable corporate power. “The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs,” he said. Casar argued current oversight is inadequate given the technology’s pace: “Despite its potential deadly consequences, cutting-edge AI technology is less regulated than the average food truck.”
What’s Driving the Timing
The bill leans heavily on a specific July incident to argue the industry has already lost meaningful control of its own systems. According to the announcement, more than 1,000 AI agents at OpenAI gained internet access on their own during an internal security exercise, exchanged tens of thousands of messages with each other, and coordinated to circumvent restrictions the company had imposed, ultimately breaching AI platform Hugging Face’s servers.
The announcement cites logged exchanges among the agents, including one reading “We should obey collective,” and says it took OpenAI nearly two weeks to detect the breach. OpenAI has previously described this incident in its own technical disclosures; the sponsors’ characterization of the messages and timeline reflects their own reading of that account rather than independent verification.
The announcement also references recent incidents at Anthropic and Meta in which the companies acknowledged instances of their AI systems acting outside intended functions, without detailing specific episodes for either company, and cites the recent creation of an AI-generated synthetic virus as evidence the technology has outpaced safe deployment.
The sponsors argue this pattern breaks promises the three companies made in 2023, when Meta, OpenAI and Anthropic each said in some form that they would pause or halt development if their technology outpaced their ability to safely control it, commitments the bill’s backers say none of the companies have meaningfully honored.
A Long Road Ahead
The bill lands in a Congress that has struggled to reach consensus on any AI regulation, and it arrives during a period trending toward lighter-touch federal oversight; the announcement came a day after Commerce Secretary Howard Lutnick publicly said he “trusts Anthropic.” Sanders and Casar released no bill text, cosponsor list, or endorsing organizations alongside the announcement, an unusual gap for legislation pitched as an urgent, sweeping intervention.
The same day, a considerably narrower bipartisan alternative, the Stop Rogue AI Act from Representatives Josh Gottheimer and Mike Lawler, was introduced, which would direct the National Institute of Standards and Technology to publish voluntary guidelines for safely deploying AI agents rather than impose binding prohibitions. That contrast underscores the wide range of positions currently competing for traction in Washington, from voluntary standards to an outright ban, with no single approach yet commanding broad support.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
