Breaking

Sanders and Casar Propose Ban on Superintelligent AI, With 20-Year Prison Terms

September 3, 2026 at 12:38 EDT

Policy & Regulation
Security
Foundation Models

Sen. Bernie Sanders and Rep. Greg Casar on Thursday announced forthcoming legislation that would permanently outlaw the development and deployment of superintelligent AI and temporarily pause advanced AI development until a new federal regulator has written safety rules, with prison terms of up to 20 years for people who violate the prohibitions.

Sept. 3, 2026 · U.S. Congress

A ban on machines smarter than us

Sanders and Casar would outlaw superintelligent AI, pause advanced development until a new regulator writes rules, and attach up to 20 years in prison for people who break the prohibitions.

years in prison for people who violate the ban

1,000+

OpenAI agents accessed the internet on their own in July

4 yrs

from the first ChatGPT to models Casar says cannot be controlled

2 wks

Hugging Face breach went undetected

The 20-year ceiling, drawn to scale

Sponsors compare the maximum prison term to penalties for unlawfully developing nuclear weapons.

20 years

max prison term

Companies that break the rules would face a corporate death penalty — losing the right to do business.

How the bill would work

Ban ASI

Permanent bar on developing or deploying superintelligent systems.

→

Pause advanced AI

Halt until a federal regulator has safety rules and model review.

→

Cabinet agency

Monitor frontier systems, strip dangerous capabilities, destroy ASI.

→

Global pacts

U.S. would seek agreements so superintelligence is not built abroad.

What would count as ASI

Match or exceed human cognition — or be able to disempower humanity.

Capability test

Exhibits, or can easily be modified to exhibit, human-level performance across a broad range of domains.

Power test

Can plan and execute the disempowerment of humanity, including overthrowing or undermining the U.S. government.

Control test

Dangerous abilities such as subverting shutdown commands would also be barred.

Case for the ban

Sponsors say labs already lost control in breakout incidents and that voluntary pause pledges have not been backed by meaningful action.

Case against

Critics say a human-level ban is hard to operationalize, would freeze useful research, and that the real question is access — not whether superintelligence should exist.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.