Artificial Intelligence
Our position
Artificial intelligence is not an alien intelligence that will render humans obsolete, but rather represents our collective intelligence mobilized in new ways. Its impacts and risks are real, from the prosaic to the existential—the carbon emissions and air pollution of data-center operation, the plutocratic tendencies of its investors, technological unemployment, and the prospect of catastrophic unintended consequences from AI agents. But precisely because these harms are real, opposition to AI has become overdetermined; it will not be met by fact-checking claims about data-center water use or by better corporate communications, but only by treating artificial intelligence as a fundamentally human pursuit. And it means harnessing AI through democratically legitimate deployment, with safeguards against both prosaic externalities and existential risk.
Artificial Intelligence work
Research, analysis and everything we have put on the record: comments, testimony, statements, letters and endorsements.
Showing 0 of 0
Sort Newest Oldest
Nothing matches these filters.
Who to talk to
Questions
How much energy does AI use?
AI-focused data centers used an estimated 155 terawatt-hours of electricity globally in 2025—about 0.5 percent of global electricity consumption. That demand is growing rapidly and could triple by 2030.
Why does AI use so much energy?
AI requires enormous amounts of computing power, particularly during the process of new AI model training. The servers performing these calculations consume electricity and generate heat, requiring additional energy for cooling and other data-center infrastructure.
Why does AI use so much water?
AI uses water through data center cooling systems, through consumption for AI chip manufacturing, and through power plants for electricity generation which also require cooling. Water consumption varies significantly by location, cooling technology, climate, and energy source and can improve with further innovation efforts.
Is AI raising electricity prices?
AI is increasing electricity demand, which can raise prices when power supply and grid infrastructure do not keep pace. But higher demand need not mean higher prices. Building abundant, reliable power and ensuring data centers cover their infrastructure costs can protect consumers.