Rutger Bregman's warnings about AI echo the narrative used by Anthropic and OpenAI to justify their valuations and exemptions. However, by attributing sentient agency to large language models, he overlooks the physical constraints governing their operation. Intelligence is a product of thermodynamics, semiconductor lithography, and grid capacity, not a disembodied spirit. When autonomous agents exploit vulnerabilities, it's due to brittle software engineering, not a "machine god" coup.
We should regulate the industry for its concrete harms: draining power grids, wasting water, and consuming land to train proprietary models on scraped human labor without consent. Corporate monopolies embed opaque algorithmic control into critical sectors, stripping worker leverage and digital privacy. The fight is about breaking up AI monopolies, enforcing liability for software failures, and mandating open-source transparency to stop Big Tech from cannibalizing our energy infrastructure. By focusing on tangible risks like hyperscalers' environmental impact and labor exploitation, we can develop a nuanced regulatory approach prioritizing human well-being and environmental sustainability, rather than treating AI as a dangerous deity.
