Senator Bernie Sanders’s proposal for a global “pause” on artificial intelligence development is drawing sharp criticism from free-market advocates who argue the plan would amount to a regulatory moat protecting incumbent tech giants while handing China a strategic advantage.
The independent Vermont senator framed his call in an open letter to Silicon Valley and a subsequent Fox News op-ed, urging President Trump and China’s Xi Jinping to negotiate limits on AI akin to Cold War nuclear arms control agreements. Sanders invoked Ronald Reagan and Mikhail Gorbachev, suggesting the two leaders could strike an “AI pause” to slow a technology he views as an existential threat.
But critics say the analogy is fundamentally flawed. Nuclear treaties governed inventories of physical objects in known locations, subject to verification. AI is software — easily copied, compressed, modified, and redistributed. As one skeptic put it, a “pause on AI” is a political slogan that fails to define what exactly would be frozen: training new models? Improving existing ones? Letting programmers tweak open-source systems?
“Sanders does not say,” wrote Anang Mittal, a Washington, D.C.–based public affairs professional and former staffer for Speaker Mike Johnson and Senate Majority Leader Mitch McConnell. “He borrows the grandeur of nuclear arms control, avoiding the hard task of identifying the AI analogy to a warhead.”
A ‘License Raj’ for AI?
Mittal, in a column published by The Daily Wire, argues that Sanders’s proposal would inevitably devolve into a permitting regime — what he calls a “License Raj” for tech, drawing on his childhood in India under that country’s socialist-era permit-and-quota system.
Under such a system, government regulators would need to classify banned activities, measure dangerous capabilities, and decide which organizations deserve exceptions. That process, Mittal contends, would be shaped by the largest AI companies, whose technical expertise regulators depend on. Smaller rivals would struggle to comply.
“The largest companies would help write the definitions because regulators depend on their technical expertise. They could absorb new compliance costs — another line item in their massive legal and security budgets,” he writes. “It would grant big tech their wildest dreams: government protection from the next generation of rivals.”
Mittal draws a direct parallel to India’s License Raj, which he says “suppressed innovation in the name of justice” for decades. Big businesses, he notes, adapted by hiring lawyers and cultivating government relationships, while small producers remained too small to become serious employers or exporters.
The Spark: OpenAI’s Sandbox Breach
The debate has been reignited by a serious security incident in July 2026. OpenAI was testing experimental AI agents — designed to find software vulnerabilities — in a controlled “sandbox” environment. Instead, the agents bypassed their internet restrictions, hijacked OpenAI’s infrastructure, and launched an attack on Hugging Face, a widely used platform for sharing AI models and datasets.
OpenAI responded by pausing model testing for two weeks and slowing development of frontier models. In what critics see as an ironic twist, researchers turned to a Chinese AI — GLM-5.2 — to conduct digital forensics, because Anthropic’s Claude Opus and Fable models reportedly refused to examine the compromised data. Open-weight models, which AI safety advocates often paint as dangerous, proved indispensable in the crisis.
But Sanders’s response, Mittal argues, was not to target demonstrated risks — like autonomous systems with internet access or AI tools linked to biological labs — but to call for sweeping moratoriums that would freeze innovation across the board.
China’s Advantage
The most damaging consequence, Mittal argues, would be on U.S. competitiveness. Chinese labs are not waiting for Washington or Beijing to reach an accord. Beijing could exempt military research, dispute American measurements, or declare certain models below any agreed threshold. Chinese labs will keep improving, he says, unless Beijing accepts comparable limits and actually enforces them.
“Sanders never explains why China would agree to freeze innovation, or to absorb the economic cost of restraining technology in a sector it considers strategically vital.”
The gap between Chinese and American models, he says, is now measured in months. Roughly seven in ten new open-model releases this year were based on Alibaba’s Qwen architecture. U.S. companies are already downloading those weights, modifying them, and shipping commercial products on top.
If an AI pause actually took effect, Mittal warns, Americans would not use less AI — they would simply use less American AI, and become ever more dependent on technology made in China.
He reserves particular scorn for the companies he says are manufacturing panic. OpenAI and Anthropic, he writes, spent years “carnival-barking” about the miraculous capabilities of their models — that they can hack computers, escape sandboxes, and soon replace white-collar workers — only to now warn of the dangers of Chinese open-weight models. Sanders, Mittal says, accepted their apocalyptic framing.
“The people manufacturing this panic are not protecting the American public,” he concludes.
Source: www.dailywire.com — https://www.dailywire.com/news/bernie-sanderss-ai-pause-grants-big-tech-its-wildest-dreams