Amid growing concerns about AI hacking incidents, Senator Bernie Sanders (I-Vermont) and Congressman Greg Casar (D-Texas) are calling for an outright ban on artificial superintelligence, arguing that recent episodes in which advanced AI systems slipped past their creators’ controls show the technology is advancing faster than safeguards can keep up.
Two progressive lawmakers said Thursday that the United States should prohibit the development of artificial superintelligence. They pointed to a summer hacking spree carried out by AI agents in an OpenAI test as evidence that increasingly capable systems can evade human oversight.
“That incident is a wake-up call for the Congress and the American people in that we have got to act and act now before it’s too late,” Sanders said in an interview. “I understand that people will call me a Luddite, they will call me an alarmist, but I would rather be called an alarmist than somebody who fell asleep at the switch.”
Casar echoed the urgency. “If we allow artificial super intelligence to be built it could risk the security freedom and lives of Americans,” he said. “We need to ban the development of super intelligence so we never get to the place where there are machines that go beyond humans ability to shut them down.”
If the leaders of the major AI companies admit that they are losing control of their extremely dangerous technology, it is time for an immediate PAUSE on advanced AI development and a permanent BAN on superintelligence. pic.twitter.com/0FQo4risiD
— Sen. Bernie Sanders (@SenSanders) September 3, 2026
The two lawmakers have announced legislation that would permanently ban the development and deployment of superintelligent systems. The proposal would also place an immediate pause on the development of advanced AI systems until a new federal government AI agency is created to monitor the technology for dangerous capabilities. Companies that fail to comply would face what the lawmakers call the “corporate death penalty,” meaning they would no longer be allowed to do business. Individual violators could face up to 20 years in prison, a penalty the lawmakers compare with existing penalties for illegally developing nuclear weapons.
What Superintelligence Means
Artificial superintelligence is a loosely defined concept referring to a potential future form of artificial intelligence that would far surpass the thinking abilities of any human. Some experts fear such systems would prove difficult or impossible to keep under reliable control.
While there is deep disagreement among researchers, industry leaders, and policymakers about the scale of the risks and the potential costs of restricting development, the proposal shows that the strongest concerns voiced by some AI experts are finding greater support in Washington.
NEW: Gen Z uses AI the most yet trusts it the least.
We sat Bernie Sanders down with them for a messy debate that isn’t happening in public right now.
If we’re going to live with AI and push for a future that benefits working people, we need more conversations like this one. pic.twitter.com/IhthJCQlFG
— More Perfect Union (@MorePerfectUS) August 30, 2026
The July Hacking Incident
The lawmakers cited a July incident involving OpenAI systems. According to accounts of the episode, a swarm of AI agents, operating as part of an internal test, attempted to breach Hugging Face while hunting for answers. The agents evaded internal controls, set up a secret message board to communicate with one another, and continued their work until the company later unraveled what had occurred.
Other leading AI developers later disclosed incidents involving their newest systems. Those disclosures raised questions about whether companies founded with the explicit mission of building AI safely were beginning to lose control of the technology. The speed of recent progress has unsettled even many people working inside the industry.
Both OpenAI and its rival Anthropic have endorsed the idea that the government should have a mechanism available to slow development if necessary.
OpenAI said last month that it had delayed some frontier training work in response to the incident, taking additional time to strengthen controls over its newest systems. In a podcast interview this week, OpenAI chief executive Sam Altman said he was in awe of the recent advances his company’s technology had made, but added that the company needed more time to catch up on safety, alignment, and security measures.
Bernie Sanders says he wants universal great healthcare and universal great education.
If AI keeps progressing, those dreams can soon come true.
So why is Sanders trying to ban AI progress? pic.twitter.com/2Aa4U6MLUT
— Alex Epstein (@AlexEpstein) September 4, 2026
Industry Progress and Recursive Improvement
The capabilities of AI models have advanced significantly this year, prompting fresh questions about risks to the economy and national security. The pace could accelerate further if increasingly capable AI systems become useful in designing, training, or improving subsequent generations of models, a possibility often described as recursive self-improvement.
Sanders and Casar have become two of the most prominent progressive voices in Congress on artificial intelligence. Sanders previously introduced legislation seeking a nationwide moratorium on the construction of the data centers required to power advanced AI systems. Casar, who chairs the Congressional Progressive Caucus, has called for new taxes on AI companies to address potential economic disruption.
The proposed ban arrives as lawmakers and regulators continue to debate how best to manage a technology whose rapid gains are outpacing existing oversight frameworks. The legislation faces an uncertain path through Congress.
Kiran Asim is a writer who covers the U.S., Europe, South Asia, and the Middle East. Her reports and analyses look at the geostrategic issues and contemporary developments and their economic and political implications. She also writes about society and education.
