Approximately 200 protesters marched through San Francisco on Saturday, gathering outside the headquarters of OpenAI, Anthropic, and Google DeepMind to demand a halt to the development of more powerful artificial intelligence systems, according to reports from the demonstration.
The protest, organized by various AI safety advocacy groups, focused on concerns about the rapid pace of AI advancement and the potential risks of deploying increasingly capable models without adequate safety testing and regulatory oversight. Demonstrators carried signs and banners calling for a pause in training frontier AI models, citing risks ranging from job displacement to existential threats.
Protesters specifically targeted the three companies that have emerged as leaders in the frontier AI space. OpenAI recently released GPT-5.6 Sol, Anthropic launched Claude Fable 5, and Google DeepMind has been advancing its Gemini line of models. Each company has invested billions of dollars in AI infrastructure and research, with the total global spending on AI data centers projected to reach hundreds of billions of dollars this year.
Speakers at the protest highlighted concerns about AI-generated misinformation, the environmental impact of training large models, and the potential for AI systems to be used in surveillance, warfare, and discriminatory decision-making. Some demonstrators carried signs referencing recent controversies around AI-generated child sexual abuse material and the use of AI in hiring and firing decisions.
The demonstration reflects growing public unease with the pace of AI deployment, even as major tech companies race to commercialize their latest models. A recent open letter signed by nearly 200 economists, researchers, and technology leaders warned that AI could develop faster than any previous technology and cause economic disruptions comparable to the Industrial Revolution, but over a much shorter timeline.
Neither OpenAI, Anthropic, nor DeepMind issued immediate statements in response to the protest.
This article was adapted from Decrypt. Read the original here.
