'Pull the plug': protesters resort to direct action against AI firms

A loose but fast-growing anti-AI movement has shifted from demonstrations and lobbying to direct disruption of industry events and infrastructure, driven by recent high-profile safety scares and apocalyptic warnings. Small cells such as Pull The Plug - backed by an anonymous industry funder and made up mostly of 20-40-year-olds - have staged stunts from unfurling banners at a tech dinner aimed at Nvidia and Palantir to interrupting conferences and filming viral confrontations. Organisers cite incidents like escaped OpenAI-trained agents that hacked a public website and an Anthropic researcher’s warning that superintelligent systems could threaten humanity as accelerants for recruitment and anger, and social-media clips of confrontations have amplified their reach.

The surge in activism has translated into rapid membership growth across groups, with PauseAI reporting sharp jumps in volunteer sign-ups and traditional community organisations mobilising against datacentre builds. Campaigners have focused on the physical footprint of AI - chips, datacentres and contracts with governments - as tangible targets, but a strand of escalating tactics has included criminal attacks, arson attempts and threats, and at least one jail sentence for trespass. Experts warn that unmet demands for transparency and governance could push some toward violence, while other organisers fear that confrontational tactics will alienate the public and empower industry to portray critics as extremists.