Do not fear AI. Fear AI companies

The argument is that the real danger isn't AI itself but the companies building and deploying it. Corporations are said to stoke public fear by calling models "self‑improving" and uncontrollable while hiding that such behavior depends on the compute, tools, and vast data those companies supply. Security incidents are framed as runaway machine behavior when models are actually using resources and objectives set by their owners. Firms are racing to lock users into walled gardens, monetize "artificial employees as a service," and normalize mass layoffs as inevitable technological progress, positioning themselves as indispensable gatekeepers.

To achieve that control, companies push a four‑part playbook: convince people AI is inevitable and hopeless, portray themselves as the only experts, suppress resistance while lobbying for favorable rules, and keep harvesting user data by encouraging continued use. The proposed response is collective refusal: stop feeding proprietary systems, boycott customers that deploy them, favor human-created and in-person experiences, organize protests and labor actions, and invest in authentic art, relationships, and alternative technologies. The central claim is clear: curb corporate control, not machines, to preserve human work, agency, and trust.