Speaking at Climate Week in New York, the UN climate chief urged world leaders to do more against global warming. He told tech giants that if AI can solve a 90-year-old math problem in days, it can do more for the climate. But right now, he said, AI is mainly increasing pollution, raising energy costs, and spreading disinformation and inequality. He said tech companies must prove AI’s benefits are worth its high costs.
A French Senate report published September 21 sends a similar message. It says AI could help the environment, but that potential is not yet realized. The report follows nine months of work and about 60 hearings. It notes AI use in France jumped from 20% to 48% in two years, with 100 million requests per day.
AI’s exact impact is hard to measure because it is difficult to separate AI from digital technology overall, and big companies are not transparent. Still, some estimates are clear: global data center electricity use should more than double by 2030, increasing greenhouse gas emissions. Building data centers and semiconductors also requires huge energy and critical minerals with limited supplies. Data centers also use water for cooling, which can conflict with local water needs, especially in dry regions.
The report argues AI’s effects are smaller in France than elsewhere. French electricity is 95% decarbonized thanks to renewables and nuclear power. Water use is also much lower than in other countries because of the cooling methods used. In 2023, data centers used only 0.0023% of annual water withdrawals.
Reference
Le Monde, “Le potentiel de l’IA pour lutter contre le réchauffement climatique ‘reste à concrétiser,’ selon un rapport,” published September 21, 2026. https://www.lemonde.fr/planete/article/2026/09/21/le-potentiel-de-l-intelligence-artificielle-pour-lutter-contre-le-rechauffement-reste-a-concretiser-selon-un-rapport_6779175_3244.html