New Bill Calls for a Pause on Advanced AI

Human eye with digital scanning interface — Photo: Rawpixel.com / Shutterstock

Senator Bernie Sanders moved to freeze the most advanced artificial intelligence while warning that tech power is concentrating into too few hands.

Story Snapshot

Sanders introduced a bill to ban artificial superintelligence and pause advanced AI until rules exist.
He proposed public ownership stakes and structural breakups to curb Big Tech dominance in AI.
President Trump dismissed AI takeover fears as a “hoax” and argued new rules are not needed.
Scholars warn that “safety” rules can entrench incumbents if regulators lean on industry expertise.

Sanders’ Push: Ban Superintelligence And Pause The Frontier

Senator Bernie Sanders announced the Ban Artificial Superintelligence Act on September 3, 2026. The bill would permanently ban the development and deployment of artificial superintelligence. It would also pause advanced AI development until a federal regulator sets safety rules, licenses, and enforcement. Sanders tied the risk to what he calls “AI oligarchs” who could build systems people cannot control. His office framed the plan as a pro-public intervention, not a pro-industry carveout.

Sanders’ recent speeches repeat a common theme. He says the richest tech leaders are “building a runaway train with no brakes.” He calls for an international treaty, like a cold-war era pact, to halt extreme systems until the world agrees on rules. Reports note he cited lab insiders who warned companies are rushing and do not know where this technology is heading. Those warnings bolster his call for a near-term pause.

Structural Remedies: Public Ownership And Corporate Separation

Beyond a pause, Sanders laid out structural ideas to cut monopoly power in AI. His American AI Sovereign Wealth Fund Act would tax half the stock of the largest AI firms and place public ownership stakes in a national fund. The plan would also force big mixed-business companies to split AI work from non-AI lines, reducing cross-subsidies and internal moats. He also proposed an Independent Commission for Democratic AI to set rules and standards over time.

The Associated Press described how the ownership plan aims to share AI gains with the public. Sanders argues this reduces control by a small group of executives and investors. He has also challenged AI leaders to stop pouring money into political groups that fight strict rules. Supporters say public stakes could counter elite capture. Critics warn such steps may replace private power with state centralization if not carefully designed.

White House Resistance: No New Guardrails Needed

President Trump rejected claims that AI could run out of control. He said fears that AI will “take over the world” are a hoax. He argued the United States already has laws to address harm, including product liability and criminal statutes. He also said strong presidential leadership is the main “guardrail” needed. The Associated Press reported his view that more rules would slow U.S. leadership and help rivals abroad.

Outside the White House, investor David Sacks echoed the case for existing tools. He urged frontier labs not to seek antitrust waivers or approval regimes in exchange for slowing down. He argued companies should accept normal liability and market discipline instead of forming a de facto cartel under a “safety” banner. That argument rejects deals that could harden industry walls against smaller rivals.

Shared Concern: Safety Rules Can Become Monopoly Moats

Researchers who study regulation warn about a trap. When technology moves fast, officials often rely on industry experts for facts and tests. That dependence can tilt rules toward the largest firms. Wide, front-end safety regimes can raise fixed costs that only incumbents can bear. Studies of artificial intelligence policy highlight this risk of regulatory capture and entry barriers if lawmakers do not build strong, open processes and competition checks.

Sanders’ bills try to solve that dilemma by attacking concentration head-on while setting strict limits on the most extreme systems. The record here shows he links safety to anti-monopoly aims, though proof is thin that “doom” messaging has already produced monopoly favors in law. The public should watch two fronts: who writes the final rules, and how those rules affect small builders’ access to data, chips, and cloud contracts. That is where power either spreads out or locks in.