Read this post on Josh’s Substack: Powering Spaceship Earth.
Environmental policy focuses on the resources used in human efforts. This is usually with a negative connotation. In the case of ChatGPT and other artificial intelligence developments, the focus has often been on energy and water.
The numbers can seem staggering. In 2023, Rolling Stone headlined a story about Microsoft’s increased water use, “AI Technology Guzzles Water: Enough to Fill 2,500 Olympic-Sized Pools.” As countless commentators have complained on TV, live streams, and in newspapers, Google’s sustainability report says they used 6.4 billion gallons of water in their offices and data centers in 2023.
In all this despairing, it’s rare that people look around at the old and familiar to ask the same questions. Andy Masley, however, has valuable context for those in despair. The resource requirements for AI development and AI’s use are far from outlandish.
Your last meal likely outweighed AI’s water consumption, especially if you ate a burger while streaming from Netflix or YouTube. In addition, lifestyle choices that Masley compares to ChatGPT use imply much greater emissions.
For further good news, energy-intensive and water-intensive industries have a natural incentive to conserve and pursue efficiency gains because doing more with less leaves more to do more with. Data centers, for example, have become incredibly efficient in energy and water use for exactly this reason. Google and all of the other big data center folks have also committed to improving the water system where they invest.
Policymakers should ensure that users pay their own way rather than banning, pausing, or otherwise hampering industries or technologies that use water or energy. Performative environmental protests or boycotts cannot solve these important problems. Instead, they require real leadership and negotiations over policy design.
Policy entrepreneurs traveling the path from here to there through the dangerous and narrow real-world policy process face a difficult and expensive trip. Advantaged groups will not simply give up their position. How shared resources like transmission lines are paid for, for example, is a thorny topic. How to identify and incentivize actions that benefit the whole of a shared system, like the electric grid, when individuals may have to pay for such actions out of their own pocket is another boondoggle. After all, environmentalists are right that plenty of industries today do not cover the costs that they impose on others.
The big picture is simply that if being green requires poverty, then it’s doomed from the start. This is Roger Pielke Jr.’s Iron Law of Climate Policy. Most voters regularly reject significant sacrifices in quality of life for environmental ends. An agenda of less is a losing proposition.
Instead, the future is about expanding what’s possible. It is about abundance, not austerity. We should throw out all of the Down Wing environmentalism and replace it with Up Wing problem-solving. As I put it in a past essay:
The challenge with water use debates is they often devolve into finger-pointing and zero-sum thinking. But the Up Wing perspective shows us a better way. Rather than banning data centers, golf courses, or data center operations and golfers ganging up on agriculture, we can embrace innovation, collaboration, and reform to ensure everyone has enough. Water users of all types are more likely to come to the table for productive conversations if we frame the topic like this instead of blaming a subset of water users.
The future isn’t about shrinking our ambitions to match limited resources—it’s about expanding what’s possible with the resources we have.
Read Masley’s whole essay for more in the same vein of environmental optimism.