Sanders and Casar propose banning the development of artificial superintelligence in the U.S. Senate

Independent Senator Bernie Sanders (Vermont) and Democratic Representative Greg Casar (Texas) are listed as the sponsors of a legislative initiative aimed at banning the development of superintelligent artificial intelligence systems in the United States; the available material does not allow verification of the proposal's official name or its exact scope. If confirmed in those terms, it would be one of the most ambitious proposals put forward in the US Congress on AI regulation, since it would not simply seek to place limits or oversight on the development of these systems, but to prohibit by law that they ever be built.

According to reports, the proposal would combine the ban on the development of artificial superintelligence with complementary oversight measures and the setting of federal safety standards. The specific breakdown of those mechanisms —deadlines, responsible agencies or possible international initiatives— is not detailed in the available material, and is therefore not specified here.

The most striking aspect of the initiative would be the penalties envisaged for those who breach it, applying both to companies and to the individual developers responsible. The nature and exact amount of those penalties is not detailed, although they are described as unusually harsh compared with other AI regulatory proposals discussed in Washington so far.

The sponsors justify the measure by pointing to recent safety incidents in advanced AI systems. The available material does not allow verification of which specific companies or models would be involved, nor confirmation of the verbatim statements attributed to the initiative's sponsors, so they are not reproduced here.

From an analytical standpoint, it is worth placing this proposal in context. Sanders and Casar belong to the most progressive wing of the US political spectrum, which makes it foreseeable that the initiative will meet strong resistance from Republicans and from much of the Democratic Party closest to the tech industry, in addition to head-on opposition from the major AI companies, which wield considerable lobbying power in Washington. Without verifiable data on its support in Congress, the proposal is better read as a political positioning and an attempt to set the terms of the debate than as a text with real prospects of moving forward any time soon.

It is also worth noting the underlying ambiguity surrounding this type of proposal: there is no agreed technical or legal definition of what exactly constitutes "artificial superintelligence", as opposed to advanced but not superintelligent AI systems. That lack of a clear dividing line could become the main practical obstacle to any attempt to turn the idea into enforceable legislation, since it would require setting capability or risk thresholds that are not standardized today, not even among experts in the field.

In any case, the proposal fits into a broader trend: growing unease, both in political circles and within parts of the industry itself, about the safety risks associated with increasingly capable AI systems. Although its approval in the short term appears unlikely given the current balance of power in Congress and the political weight of the tech industry, initiatives of this kind help push the public debate toward demanding greater safety guarantees, and could serve as a reference for future regulatory proposals that are less extreme but inspired by the same concerns.