🤖⚡ Does AI use too much energy? A Quick Q&A with … policy analyst James Broughel

'AI, following crypto, could well be the next electricity scapegoat.'

Quote of the Issue

“Energy is the only universal currency: one of its many forms must be transformed to get anything done.” - Vaclav Smil, Energy and Civilization

---

The Conservative Futurist: How To Create the Sci-Fi World We Were Promised

“With groundbreaking ideas and sharp analysis, Pethokoukis provides a detailed roadmap to a fantastic future filled with incredible progress and prosperity that is both optimistic and realistic.”

---

The Essay

🤖⚡ Does AI use too much energy? A Quick Q&A with … policy analyst James Broughel

The promise of recent advances in AI is tremendous, from higher worker productivity to turbocharging scientific progress. But there are potential downsides, too: job loss, disinformation and deepfakes, and existential risk. Some critics also think AI uses an unacceptable amount of energy. From a recent NYT piece on the issue:

OpenAI’s ChatGPT exploded onto the scene nearly a year ago, reaching an estimated 100 million users in two months and setting off an A.I. boom. Behind the scenes, the technology relies on thousands of specialized computer chips. And in the coming years, they could consume immense amounts of electricity.

A peer-reviewed analysis published Tuesday lays out some early estimates. In a middle-ground scenario, by 2027 A.I. servers could use between 85 to 134 terawatt hours (Twh) annually. That’s similar to what Argentina, the Netherlands and Sweden each use in a year, and is about 0.5 percent of the world's current electricity use.

And this from the Wall Street Journal:

Since 2010, power consumption for data centers has remained nearly flat, as a proportion of global electricity production, at about 1% of that figure, according to the International Energy Agency. But the rapid adoption of AI could represent a sea change in how much electricity is required to run the internet—specifically, the data centers that comprise the cloud, and make possible all the digital services we rely on.

This means the AI industry is poised to run the equivalent of a planetary-scale experiment, according to experts both outside and inside the industry. This has some of them wringing their hands, while energy suppliers are practically salivating at the expected increase in demand.

I read those stories and think: “One more reason to push for new clean energy sources!” But there are Down Wingers and degrowth environmentalists who see AI energy consumption — and consumption, more broadly — as a reason to constrain the technology. I mean, of course they do.

To find out more about where the AI energy debate might take us, I asked economist James Broughel, who recently wrote a Forbes column on the issue, a few quick questions. Broughel is a senior fellow at the Competitive Enterprise Institute, where he concentrates on the economic impact of regulation. In addition to his regular Forbes column, he is the author of Regulation and Economic Growth: Applying Economic Theory to Public Policy.

1/ Do you think there is a real risk of restrictions on AI electricity usage becoming an active political issue?

Absolutely. In fact, it is already starting to happen as evidenced by news stories and commentary online.

We have seen a version of this dynamic play out previously with the debates surrounding cryptocurrency mining and its electricity demands. When an industry lacks trust and familiarity among the public, it can become vulnerable to politically-motivated attacks, in some cases using tangential issues like electricity as a focal point.

AI, following crypto, could well be the next electricity scapegoat. I should caveat this by saying there are reasonable concerns regarding AI’s environmental footprint, especially given the compute-intensive nature of the technology. However, there is also a risk that reasonable concerns get weaponized for political ends. We saw this with Operation Chokepoint, which was an effort by the Obama administration to go after unpopular companies like gun dealers and payday lenders by targeting the banks that serve these industries. Now we are seeing a focus on access to electricity rather than to the financial system.

When you combine unease about new technologies with concerns about climate change, it is almost inevitable that AI electricity use will become a focus of politics in the coming months and years.

2/ What is the case against worrying how much energy AI uses?

Leading technology companies are taking steps to procure electricity from renewable and other low-emission sources. For example, Microsoft cut a deal with Helion Energy last year to eventually purchase electricity from fusion power sources, while cryptocurrency mining operations are also striking deals with nuclear and renewable power companies, or locating facilities near places like Niagara Falls to take advantage of abundant hydropower. Apple has also secured commitments from manufacturers to use cleaner energy.

It’s plausible the tech industry’s high electricity demand may not be something to be feared at all. The industry’s heavy spending could actually encourage greater adoption of renewables and next-generation nuclear technology by creating a reliable customer for what are currently fairly risky new power sources.

Given the tech industry is well aware of its electricity footprint and recognizes that this could be a source of substantial criticism, we should expect Big Tech to prioritize improving energy efficiency and purchasing electricity from cleaner sources.