Activists with the group Stop the AI Race marched between the San Francisco offices of OpenAI, Anthropic and Google DeepMind on Saturday, the movement’s second demonstration in four months, renewing a demand that the three companies’ chief executives publicly commit to pausing frontier AI development if competitors do the same.
The march followed a route from OpenAI’s Mission Bay headquarters to Anthropic’s offices and on to Google, a reversal of March’s route, which began at Anthropic before moving to OpenAI and xAI. Organizers said the change in the third company visited, Google DeepMind in place of xAI, reflected an effort to press Google chief executive Demis Hassabis directly after previous remarks he made on the subject.
The group’s core demand has remained unchanged since its first march in March, when roughly 200 participants walked between the same companies’ offices: that OpenAI’s Sam Altman, Anthropic’s Dario Amodei and Google DeepMind’s Hassabis each publicly agree to pause development of more advanced frontier models if every other major AI lab in the world verifiably agrees to do the same.
Organizers said the companies have moved only partway toward that commitment since March. According to the group, Anthropic has stated it “would slow down or temporarily pause” development if other labs verifiably did so as well, while OpenAI has said it expects international coordination, including slowing frontier development, to become more important going forward. “These are first steps, but ‘expects’ isn’t a commitment,” the group said in materials promoting Saturday’s march, adding that it wants specific commitments from each CEO along with a concrete verification regime.
David Krueger, an AI professor at the University of Montreal and co-author of research with AI safety researchers Yoshua Bengio and Geoffrey Hinton, was among the speakers at Saturday’s demonstration. Krueger has said AI companies are “taking away” workers’ livelihoods and power in exchange for promises the industry has not kept.
The demonstration is the latest action from Stop the AI Race, a movement founded by filmmaker and former AI safety researcher Michael Trazzi, who led an 18-day hunger strike outside Google DeepMind’s London offices in September 2025. Hassabis said at the time he would be open to a conditional pause but described international coordination as the key obstacle. Anthropic separately dropped its own Responsible Scaling Policy commitment to pause development of dangerously capable models in February, replacing it with a revised, non-binding version.
The march comes as the Trump administration pushes to establish a national AI regulatory framework intended to preempt individual state AI laws, a move the administration has framed around maintaining U.S. competitiveness against China in frontier AI development. Organizers of Saturday’s march described the demonstration as peaceful and said they welcome participation from employees of the companies involved. OpenAI, Anthropic and Google DeepMind did not immediately respond to requests for comment on Saturday’s demonstration.
