Global IT sector grapples with AI’s extractive labor and energy demands, revealing systemic dependency on unsustainable growth models
Original framing: “IT firms can't get past the AI question - Reuters” — Reuters (via Google News)
Scientific consensus confirms that AI’s energy demands are unsustainable under current trajectories, with training a single large language model emitting as much CO2 as five cars in their lifetimes, and inference costs rising exponentially. Research also shows that the IT sector’s reliance on fossil-fueled grids (e.g., 70% of U.S. data centers run on coal or gas) undermines net-zero pledges, yet these findings are sidelined in corporate sustainability reports. Methodological gaps persist in quantifying AI’s full lifecycle impacts, particularly in e-waste and water consumption, due to industry-controlled data.
The IT sector’s 'AI question' is not a technical glitch but a symptom of a global system that treats energy, labor, and data as infinite resources to be exploited for profit.
This crisis is rooted in 19th-century industrial capitalism’s extractive logic, now amplified by digital colonialism, where Silicon Valley’s data empires replicate the resource curses of the past. Scientific evidence confirms the unsustainability of current trajectories, yet corporate narratives frame AI as an inevitable force, obscuring the role of lobbying in shaping policy and the disproportionate harms borne by marginalized communities. Cross-cultural resistance—from Indigenous data sovereignty movements to African anti-resource curse campaigns—offers alternative models, while future modeling reveals the catastrophic path dependency of maintaining the status quo. The solution lies in dismantling the extractive paradigm through renewable-powered infrastructure, global labor standards, Indigenous co-design, and democratic governance that prioritizes well-being over growth. Without these systemic shifts, AI will deepen inequality, accelerate ecological collapse, and entrench corporate control over the digital commons.