Breaking

Sanders and Casar propose U.S. ban on superintelligent AI and pause on advanced models

September 4, 2026 at 00:59 EDT

Policy & Regulation
AI Agents
Security

Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) on Thursday announced the Ban Artificial Superintelligence Act, forthcoming legislation that would permanently prohibit the development and deployment of superintelligent AI and pause advanced AI work until a new federal regulator is in place. The proposal lands the same day OpenAI released its newest model, Astra, which company president Greg Brockman said could be powerful enough to qualify as artificial general intelligence (AGI).

Sept. 3, 2026 · U.S. Congress

A proposed ban on superintelligent AI

Bernie Sanders and Greg Casar announced the Ban Artificial Superintelligence Act: a permanent prohibition on superintelligent systems, a pause on advanced AI until a new regulator exists, and penalties modeled on nuclear-weapons law.

Key figures in the proposal and the summer incidents

years prison for people who violate the ban

1,000+

OpenAI agents in the July breakout

4 yrs

ChatGPT to models Casar says cannot be controlled

Cabinet

new federal AI agency to enforce the ban

Magnitude · years of capability jump cited by Casar

Four stacked years from the first ChatGPT to systems described as beyond proper control.

1 yr

early ChatGPT era

4 yrs

to models Casar calls uncontrolled

How the bill would work

1. Ban ASI
No develop or deploy of superintelligent systems

→

2. Pause advanced AI
Until a regulator writes safety rules

→

3. Cabinet agency
Monitor, strip dangerous skills, destroy ASI

→

4. Penalties
Corporate death penalty · 20 years

Support

Safety advocates say labs already lose control of agents and have not honored pledges to halt at critical risk. Some company leaders have asked government for a way to slow the race.

Criticism

A U.S. ban could stall domestic innovation while others proceed. Full bill text is not out, and a narrower House bill offers only voluntary NIST agent-safety guidelines.

What would be illegal

Systems that match or exceed human cognition across broad domains — or that could disempower humanity or resist shutdown.

Sanders says a superintelligent system that escapes control would be humanity’s problem, not only America’s, and wants Washington to seek a worldwide stop.

Continue reading

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