Protesters urge Trump and Xi to negotiate a treaty restricting advanced AI
The coordinated demonstrations called for verifiable limits on superintelligent AI, while participants differed over whether existing models should also be withdrawn.
Loading page…
The coordinated demonstrations called for verifiable limits on superintelligent AI, while participants differed over whether existing models should also be withdrawn.
Listen to this story
The proposed treaty would make any pause reciprocal and verifiable: set limits on training computation and track AI chips, rather than ban AI outright. Its stated aim is to stop premature development of superintelligent systems until it can proceed without the risks outlined in the proposal. The demonstrations also surfaced differing views about whether restrictions should cover only future development or extend to existing models and data-center construction; the article reports calls for negotiation, not a resulting agreement.
More than 100 people marched in San Francisco on November 14, 2026; organizers reported coordinated demonstrations in about 20 cities.
The proposal, developed by the Machine Intelligence Research Institute’s Technical Governance Team, would use FLOP thresholds and chip tracking to verify limits.
Pause AI supports pausing further advanced development without rolling back existing models; one Bay Area coordinator personally advocated withdrawing some systems.
More than 100 demonstrators marched through San Francisco on November 14, 2026, calling on Donald Trump and Xi Jinping to negotiate an international treaty restricting advanced AI development. Organized by Stop The AI Race, the protest sought government action on a demand that goes beyond voluntary company restraint: verifiable limits on the pursuit of superintelligent AI.
The crowd gathered outside OpenAI’s offices on Third Street before marching about two miles to the Veterans Memorial. Organizers displayed “AI TREATY NOW” banners and a large papier-mache sculpture of clasped hands, one bearing the U.S. flag and the other China’s. Members of Pause AI and Berkeley-based Stop AI also attended, according to 48 Hills.
The San Francisco action was part of a coordinated international campaign. Latin Times reported demonstrations in approximately 20 cities, including Washington, New York, Paris, Montreal and Berlin. Protesters wanted the American and Chinese leaders to take up AI restrictions at the APEC forum in Shenzhen, China.
The movement’s formal proposal is not a permanent ban on artificial intelligence. As described by Latin Times, it would halt the premature development of artificial superintelligence—systems more cognitively capable than humans—until development could proceed without the risks identified in the proposal. Its framework was developed by the Machine Intelligence Research Institute’s Technical Governance Team.
The international scope addresses a problem participants described as competitive pressure. Former software engineer Zack M. Davis told AFP that if Anthropic stopped alone, OpenAI could move ahead; if the United States stopped alone, China could do the same, according to Latin Times. The proposed answer is a shared restriction rather than a unilateral retreat.
We believe that the right move is for our whole society to pace the frontier of AI development in a broad international treaty
Thomas White, Stop The AI Race volunteer, speaking to 48 Hills
Participants did not all draw the same boundary around acceptable AI. Gregory John Ward, Pause AI’s Bay Area coordinator, said his organization supports an international pause on advanced development without rolling back existing models. His personal position went further: some released models should be withdrawn unless they are provably safe.
Ward linked that stricter position to trust, arguing that putting untrustworthy AI in charge of important institutions could undermine society. That was his argument for withdrawing some existing systems, distinct from the organization’s stated policy of pausing further advanced development while leaving current models in place.
Other speakers brought infrastructure and political power into the protest. Food and Water Watch managing director Mark Schlosberg called for a moratorium on data-center construction. He argued that the buildout was polluting waterways, displacing jobs and enabling an unwanted superintelligence, presenting construction limits as another way to slow development.
Attendee Jeremy Ornstein, who was affiliated with Stop The AI Race, urged protesters to confront recently elected politicians over promises to curb AI development. He compared U.S.-China competition to the nuclear arms race and argued that a treaty could work if each country could verify that the other was slowing down.
At the march’s endpoint, movement founder Michaël Trazzi asked attendees to undertake hunger strikes to pressure AI companies to slow development. He had conducted one himself the previous year. Alongside the appeal to heads of state, his closing request asked individual protesters to escalate their own commitment to the campaign.
Loading discussion...
Join the conversation
Explain where you would draw the line.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.