'Is it really necessary to generate another image?': UN scientist explains how everyday people can limit AI's environmental impact

The environmental cost of artificial intelligence is no longer a niche concern—it’s a growing crisis hiding in plain sight. Kaveh Madani’s insights from the UN report underscore a paradox: while AI promises efficiency and progress, its insatiable demand for energy is accelerating the climate crisis it was designed to help solve. This isn’t just about data centers humming in isolation; it’s about a technology whose carbon footprint is expanding faster than renewable energy can offset it. For everyday users, the question isn’t whether AI will reshape industries but whether society can afford its ecological toll. What many overlook is the scale of AI’s hidden infrastructure. Training a single advanced model can consume as much electricity as a small city, with inference—the process of using the model—adding another layer of energy drain. Yet the real issue lies in the compounding effect: as AI integrates deeper into daily life—from smartphones to smart cities—its energy demands spiral. Madani’s emphasis on user behavior hints at a critical insight: the average person’s role in this isn’t passive. Every unnecessary query, every redundant image generated for convenience, feeds a system that’s already operating at unsustainable levels. The next phase of this debate will likely focus on regulation versus personal accountability. Will governments cap AI’s energy usage, or will consumers bear the burden of change? Already, some tech giants are exploring greener alternatives, but these efforts remain piecemeal. Meanwhile, the public’s reliance on AI tools shows no signs of slowing, creating a tension between innovation and sustainability that demands urgent resolution. This issue also reflects a broader tension in the tech industry: the race to deploy AI at any cost versus the need for deliberate, measured progress. As AI becomes more embedded in critical systems—healthcare, transportation, finance—the stakes grow higher. The question isn’t just about reducing energy use today but ensuring that the technology of tomorrow doesn’t inherit an unfixable environmental debt. The conversation Madani’s report sparks is overdue, but its outcome will define whether AI becomes part of the solution—or another driver of the problem.