Simon Stiell, executive secretary of the United Nations Framework Convention on Climate Change, told world leaders at Climate Week in New York on Monday that artificial intelligence is currently worsening climate problems rather than solving them. His speech coincided with a French Sénat report that found AI's environmental potential "has yet to be realized," putting direct pressure on technology companies to account for the technology's growing energy footprint.
Stiell addressed the "tech titans" directly. "If AI can solve a 90-year-old math problem in days, it can do more on key climate challenges," he said. He argued that AI is mostly "heating pollution from coal, oil and gas, while ratcheting up energy costs for households and businesses" and "fueling disinformation, social divisions and even larger inequalities of wealth."
The gap between promise and reality
The French Sénat's committee on AI's environmental impact, led by senators Guillaume Chevrollier, Damien Michallet, and Simon Uzenat, published its findings the same day. The report echoes Stiell's assessment: AI could serve environmental goals, but that potential remains untapped. The senators stopped short of dismissing AI's climate applications entirely, instead framing the issue as one of unrealized capability.
Data centers, which power AI workloads, require enormous volumes of chilled water and electricity. A photo from Noyal-sur-Vilaine in France, showing the chilled water pipes of a data center cooling system, accompanied coverage of the week's events. These physical infrastructure demands are rarely visible to users of AI tools, but they represent a direct line from digital activity to resource consumption.
Stiell's demand for accountability
Stiell did not call for abandoning AI. He called for a shift in who benefits from it. "Tech titans need to start showing why the benefits of AI outweigh its skyrocketing costs - for the many, not just the tiny few," he said. His framing ties AI's climate costs directly to questions of equity and corporate responsibility.
The combined messages from the UN and the French Sénat signal a growing impatience with vague claims about AI's climate potential. Both institutions want measurable results, not promises. For professionals who build or manage the physical world, the tension between AI's computational demands and its potential optimization gains is no longer theoretical.
Why this matters for real estate, construction, and research
Data center construction is booming, and those facilities need land, power, water, and cooling infrastructure. Real estate and construction professionals are on the front line of siting and building these projects, often in regions already stressed for water and grid capacity. The Sénat report and Stiell's remarks suggest that regulatory scrutiny of data center resource use will intensify. Professionals who understand the energy and water demands of AI infrastructure-and who can plan for stricter efficiency standards-will be better positioned as projects face tougher environmental reviews. For researchers, the gap Stiell identified between AI's climate promise and its current reality points to a clear need: developing and validating AI applications that deliver measurable emissions reductions, not just computational speed.
Your membership also unlocks:
