Al Gore Says AI's Real Danger Isn't Data Center Emissions — It's What the Industry Itself Is Warning About
In a new TechCrunch interview, the former vice president downplays AI's energy footprint as his top concern and points instead to the existential risks AI leaders themselves keep flagging.
Al Gore is not lying awake over the electricity bill for AI data centers. In an interview with TechCrunch, the former vice president said the emissions tied to the computing buildout rank well below a different worry: the warnings coming from the people building the technology. Gore's argument is less about carbon accounting than about which risk deserves attention first, and he is siding with the industry's own doomsayers.
That framing cuts against the current center of gravity in climate policy. Data centers have become a favorite target for regulators and utilities, with grid operators projecting sharp load growth from AI training and inference, and several tech companies citing power demand as they walk back clean-energy targets. Gore's position suggests he sees that debate as a manageable engineering and procurement problem, not an existential one.
The more serious signal, in his view, is the steady stream of cautions from AI executives and researchers themselves. Figures across the leading labs have signed statements comparing unaligned or misused AI to pandemics and nuclear war, and some have put the odds of catastrophic outcomes high enough to justify dedicated safety work. When the people with the most information about a system are the ones raising alarms, Gore appears to treat that as the load-bearing evidence.
There is a political logic to the choice as well. Gore has spent decades arguing that climate change is an urgent, solvable problem, and he has watched that case get tangled in disputes over individual technologies and short-term costs. Casting AI's energy use as a solvable side issue lets him keep the climate message intact while acknowledging a newer set of risks that governments have barely begun to regulate.
The practical question is what follows from the distinction. If data center emissions are a second-order problem, the policy energy now aimed at siting, water use and utility rates may be pointed at the wrong target, while safety testing, model evaluation and deployment rules remain thin. Gore's comments add a prominent voice to the camp arguing that the sequencing is backwards.
What to watch is whether his framing changes the terms of the debate or simply splits it. AI companies are under pressure on both fronts — from regulators scrutinizing their power consumption and from critics demanding proof their systems are safe — and they have an incentive to treat the easier problem as the real one. Gore's intervention makes that trade-off harder to hide.