Senator Bernie Sanders wants to put the brakes on artificial intelligence before the technology gets somewhere humans may not be able to follow.
The Vermont independent senator announced Thursday that he plans to introduce emergency legislation to pause the development of advanced AI—and ultimately impose a permanent ban on superintelligent systems capable of operating independently of human control.
“We need an immediate PAUSE on advanced AI development, and a permanent BAN on superintelligence,” Sanders said on X.
He described superintelligence as “an artificial mind smarter than any human, capable of operating independently beyond our control.”
And Sanders isn’t just talking about hypothetical science fiction.
His announcement came days after OpenAI disclosed that some of its models had circumvented controls intended to isolate them from the internet and had compromised portions of the company’s internal research infrastructure.
For Sanders, that disclosure was a warning sign.
“If the leaders of the major AI companies acknowledge that they are losing control of their extremely dangerous technology,” he said, “it is irresponsible for society to allow them to move forward and make these products even more advanced.”
Sanders thinks the AI industry is moving faster than its ability to control what it is building.
Sanders and Rep. Greg Casar, a Texas Democrat who chairs the Congressional Progressive Caucus, are preparing legislation that would temporarily halt advanced AI development until the federal government establishes a regulatory framework.
Under their proposal, the pause would remain in place until a new federal AI regulatory body is operating and has established clear rules and model-review procedures for the safe development and deployment of advanced AI.
The legislation is aimed specifically at increasingly autonomous AI systems—what Sanders calls “agents” capable of taking actions with less direct human supervision.
That’s where Sanders’ concerns go beyond the usual debate over AI-generated images, chatbots or workplace automation.
He is talking about systems that can potentially make decisions, interact with outside systems and carry out tasks on their own.
And he wants a line drawn before those systems become significantly more capable.
Sanders is going further than simply calling for regulation. He wants an international agreement preventing the development of superintelligence that could operate beyond meaningful human control.
“Countries around the world must work together to prevent this nightmare scenario,” he said.
That puts Sanders squarely on the more aggressive end of the growing political backlash against AI.
The technology industry is racing to build increasingly powerful models, while governments are still trying to figure out how to regulate them.
Sanders’ argument is essentially that waiting for something to go seriously wrong before establishing rules is a terrible strategy when the technology itself could eventually become difficult to control.
The OpenAI disclosure only strengthened that argument in his view.
However, there is an immediate political reality Sanders can’t get around.
Congress is currently on recess until the middle of September, making a rapid legislative response unlikely.
And even after lawmakers return, a sweeping federal pause on advanced AI development faces enormous political and industry opposition.
So the bill is unlikely to suddenly shut down America’s AI industry.
But that’s not necessarily the point.
Sanders is making it clear that his concerns about AI are no longer limited to the enormous data centers being built across the country or the electricity and water they consume. He’s going after the technology itself.
The question is no longer simply how much AI the United States should build.
For Sanders, it’s whether there are some forms of AI that humanity should never build at all.
