Sanders Turns AI Safety Into a US-China Power Contest
Sanders is trying to shift AI policy from rivalry to risk reduction, but Trump officials and China hawks still control the leverage.
Bernie Sanders is not trying to win the AI race; he is trying to change the race’s premise. In remarks reported by 
The Hill, the Vermont senator called artificial intelligence a “runaway train” and argued that the U.S. and China should cooperate to prevent a “cataclysmic development,” even as Washington increasingly treats AI as a strategic competition. That is a minority view in US Politics and in the broader Washington consensus — but it is a deliberate one.
Sanders is fighting the frame, not the industry
Sanders’ leverage is political, not institutional. He does not control export controls, chip sales, or federal AI rules. The people who do — the Trump White House, Treasury Secretary Scott Bessent, and congressional China hawks — are pushing the opposite message: America must set the standard, and Chinese involvement is a risk, not a remedy. 
Responsible Statecraft reported that Bessent blasted Sanders for inviting Chinese scientists to a Capitol Hill event, accusing him of letting “foreign nationals” shape U.S. regulation.
That clash matters because AI policy is currently being written as a sovereignty contest. Sanders is trying to recast it as a shared hazard, closer to nuclear risk than to trade policy. The nuclear analogy is not decorative; it is the only model that justifies talks with Beijing when the dominant U.S. instinct is to deny China any say over global standards.
Washington is betting on competition; Sanders is betting on fear
The political environment is stacked against him. Reuters reported on April 21 that the Trump administration’s China policy has been drifting toward managed competition, with officials emphasizing leverage, tariffs, and technology restrictions rather than durable cooperation; the same Reuters report said Trump was preparing for a mid-May meeting with Xi Jinping. Separately, 
NPR/AP reported that the administration is cracking down on Chinese firms it says are extracting capabilities from U.S. AI models, while also warning that the U.S.-China performance gap in top models has effectively closed.
That is the central problem for Sanders’ argument. The more Washington believes the gap has narrowed, the less appetite there is for shared guardrails — because policymakers start seeing cooperation as a transfer of advantage. Beijing, for its part, can welcome dialogue, but it is unlikely to accept a framework that limits its own industrial and military upside while the U.S. keeps scaling.
What this changes
Sanders may not move the executive branch, but he is shaping the left edge of the AI debate inside Congress. That can still matter if the risk narrative spreads beyond progressives and into mainstream oversight hearings, especially if child safety, labor displacement, or model misuse becomes harder to ignore. It also gives Chinese academics and safety advocates a rare U.S. platform, which is why the event drew such a sharp reaction from the administration.
The next test is the May 14-15 Trump-Xi meeting, and whether AI safety appears in the joint readout. If it does, Sanders has opened a lane. If it doesn’t, this remains what Washington usually makes of AI risk: a warning, not a policy.
