WASHINGTON, September 14 — Senator Bernie Sanders has introduced legislation that would ban the development of superintelligent artificial intelligence, attaching criminal penalties of up to 20 years in prison for developers who breach the prohibition. The bill defines superintelligent AI as a system that surpasses human cognitive performance across virtually all economically valuable tasks. No known system has reached that threshold.
The proposal lands amid a global surge in AI investment and deployment, with large language models from OpenAI, Google and Anthropic already in wide circulation.
A familiar target
Sanders, the independent from Vermont, was first elected to the Senate in 2006 after serving as mayor of Burlington and as a US representative. He has spent much of his career challenging corporate power, through his 2016 and 2020 presidential campaigns and his advocacy for Medicare for All. He has previously raised concerns about AI’s effect on workers and on democratic institutions.
The new bill extends that argument to the technology itself. The penalties would rank among the harshest ever proposed for technology development.
They could fall on executives and engineers at leading AI firms, among them OpenAI, Google DeepMind, Anthropic and Meta. The legislation reflects a widening unease among policymakers and technologists about existential risk from advanced AI. That debate was sharpened by the 2023 open letter from the Future of Life Institute, which called for a six-month pause on large AI experiments and drew signatures from figures including Elon Musk and Steve Wozniak.
An uncertain path
Prospects in a divided Congress are uncertain. The tech industry and lawmakers who favour a more permissive regulatory approach are likely to oppose it strongly. The bill also intersects with broader international efforts, including the European Union’s AI Act, which classifies AI systems by risk level.
If enacted, it would be the most restrictive AI law in the world. Whether it moves at all is another matter.
For all the noise surrounding AI regulation, the salient point is that no committee vote has been scheduled and the bill’s text has not been released in full. Sanders’s office has issued no detailed statement beyond the bill’s introduction. Watch for the text itself, and for whether the tech firms named in the proposal choose to lobby against it in public or in private.
