A Data for Progress survey of more than 1,300 likely U.S. voters found that 68% support the Ban Artificial Superintelligence Act, introduced by Sen. Bernie Sanders and Rep. Greg Casar, against just 25% opposition.
The bill has two core provisions. First, a permanent ban on AI systems that surpass human cognition and can resist human attempts to shut them down. Second, a temporary pause on advanced AI development until a new federal regulatory body establishes safety rules and processes.
What the Bill Would Do
Sanders and Casar framed the legislation as a bid to stop what they call “AI oligarchs” from building machines humans cannot control.
The bill’s definition of “superintelligent AI” is deliberate and specific. It covers systems capable of outpacing human intelligence and resisting shutdown, not the chatbots and image generators you already use every day.
The partisan breakdown in the Data for Progress poll is notably uniform across party lines:
- 68% overall support, 25% oppose (1,300-plus likely voters surveyed)
- 72% of Democrats support the bill
- 70% of independents and third-party voters support it
- 63% of Republicans back it
- A separate June 2026 Data for Progress poll found 63% of voters support a one-year moratorium on new AI data center construction
Researchers Sound the Alarm
The poll follows a cluster of unusually blunt public departures from the labs most responsible for pushing AI capabilities forward.
Jacob Coxon resigned from Anthropic after working as an AI safety and pretraining researcher at both Anthropic and OpenAI. In a public statement reported by CNBC, Coxon said: “Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.”
Evan Hubinger, an alignment science lead at Anthropic, backed that assessment. According to CNBC, Hubinger personally estimates the probability of AI killing all humans within a decade at greater than 10%. He also acknowledged that Anthropic “does not yet have a plan to solve alignment for superintelligence.”
Alex Turner, a former research scientist at Google DeepMind, stated that many researchers believe they are building something that could kill everyone on the planet, according to Common Dreams. Jason Wolfe, a researcher at OpenAI, warned that at the current “frankly terrifying pace,” humanity would be “quite lucky” to navigate the narrow path between catastrophic outcomes, per Common Dreams.
Jonathan Richard Schwarz, a former senior research scientist at Google DeepMind, declined offers from other major AI labs entirely. He cited “severe concerns about the concentration of power these labs represent.”
The Political Context
If you follow federal tech policy, this level of cross-partisan agreement is rare territory in modern U.S. politics.
Sanders has pointed directly at these researchers as evidence for federal action. He stated that “the very people building this technology admit that it could threaten the future of humanity,” according to Common Dreams.
This is not the first call for a pause. Open letters in 2023 and 2024 urged AI labs to slow down; none produced binding policy. The Sanders-Casar bill represents one of the first serious congressional attempts to codify both a development pause and a permanent ban on superintelligence.
The available research does not yet document detailed industry responses to this specific bill. Critics of broad AI restrictions have generally cited innovation costs and global competitiveness, but how Anthropic, OpenAI, or Google DeepMind will engage with this legislation is not yet on the record.
Sixty-eight percent is the kind of number that moves campaigns. Michigan Democratic congressional candidate William Lawrence is already running explicitly against uncontrolled AI development. The more pressing question for AI companies may not be whether federal regulation is coming, but how fast.
