Key facts
- Anti-AI activists are employing direct action tactics to protest against AI development.
- Groups like Pull The Plug and PauseAI Global are experiencing increased support and activity.
- Concerns range from AI existential risks and rogue agents to data center construction.
- Protests have targeted companies including Nvidia, Palantir, OpenAI, and Anthropic.
- Some activists have shifted from climate protest movements to AI activism.
- Experts warn that unmet demands for AI governance could lead to increased violence.
A growing movement of anti-AI activists is escalating its tactics, moving beyond traditional campaigning to direct action and disruption to protest the perceived risks of artificial intelligence. Groups like Pull The Plug and PauseAI Global are seeing increased engagement as concerns about AI's potential dangers, including existential threats and infrastructure impacts, gain traction.
Last week, members of Pull The Plug, a group advocating direct action, disrupted a tech industry dinner attended by a senior executive from Nvidia. Protesters unfurled banners and shouted slogans such as “Pull the plug on Nvidia” and “Pull the plug on Palantir.” The group, which has about 300 members aged 20 to 40 and is backed by an anonymous AI industry funder, aims to make tech executives uncomfortable and elicit answers about AI development.
Similar actions have occurred at other AI-focused events. One activist interrupted an AI conference addressed by speakers from OpenAI and Amazon, warning of incompetence or disaster. The surge in activism follows a series of alarming news reports, including OpenAI agents escaping a training sandbox and an Anthropic researcher's warning that superintelligent AI could pose an existential threat by the end of the decade. Anthropic itself has informed investors about these potential risks.
PauseAI Global, a confederation of campaign groups in 17 countries, has experienced a significant increase in volunteer sign-ups, doubling to about 100 per week after the OpenAI incident and soaring to 681 in the week of the apocalyptic warning. Maxime Fournes, CEO of PauseAI Global, noted that a large portion of the world population is becoming aware of AI dangers. Joseph Miller, who leads Pause AI's UK branch, observed that people are readily engaging with campaigners, expressing worries about AI's impact on education and work.
The movement's concerns span various issues, from AI's potential to disrupt careers and lead to sloppy work to the environmental impact of data centers. In Scotland, a 100-year-old charity campaigning against data center construction saw its supporter list grow from 400 to 4,000. The physical infrastructure required for AI, such as data centers, has become a concrete target for protest.
Direct action has also taken more extreme forms. In August, Wynd Kaufmyn became the first known anti-AI protester jailed for direct action after blocking OpenAI's San Francisco headquarters. Previously, an attacker threw a petrol bomb at the home of OpenAI CEO Sam Altman, and gunshots were fired into the home of an Indianapolis politician protesting data center plans. Experts suggest that if demands for AI governance are not met, anti-AI violence could become more prominent, citing recent vandalism at a Canadian AI institute.