How tech companies could shrink AI’s climate footprint : Short Wave

AI is rapidly transforming how we live, work, and communicate. But can we undergo that transformation without destroying the environment?

Hiroshi Watanabe/Getty Images

hide caption

toggle caption

Hiroshi Watanabe/Getty Images

AI is rapidly transforming how we live, work, and communicate. But can we undergo that transformation without destroying the environment?

Hiroshi Watanabe/Getty Images

In 2018, computer scientist Sasha Luccioni was an AI researcher for Morgan Stanley — and couldn’t shake this existential worry.

“I essentially was getting more and more climate anxiety. I was really feeling this profound disconnect between my job and my values and the things that I cared about,” Luccioni told NPR.

So Luccioni quit her job.

Now the Climate Lead at Hugging Face, an online community for AI developers to share models and datasets, Luccioni is part of a growing movement to make AI more environmentally sustainable.

One solution? Less artificial intelligence.

It’s not the only solution. In her 2023 TED talk, Luccioni encouraged the adoption of small AI models. Small language models (SLMs) have far fewer parameters and require much less energy than general-purpose large language models (LLMs), such as ChatGPT.

“Nowadays, more companies are like, ‘For our intents and purposes, we want to summarize PDFs.’ You don’t need a general purpose model for that. You can use a model that is task specific and a lot smaller and a lot cheaper,” Luccioni told NPR.

As AI models have grown in size, so have the energy required to run and maintain their infrastructure. A 2024 report by Lawrence Berkeley National Laboratory forecast that by 2028, U.S. data centers could consume as much as 12% of the nation’s electricity.

The same year, Google reported that their greenhouse gas emissions increased by almost 50% in the last five years, due in part to the AI boom. In the U.S., 20 new large data centers are slated for construction through the private joint venture Stargate.

Meanwhile, Google, Microsoft and Meta have also pledged to reach at least net-zero carbon emissions by 2030. Amazon has set their net-zero deadline for 2040. (All four companies are financials supporters of NPR. Amazon also pays to distribute some of NPR’s content.)

Two additional ways tech companies are seeking to offset their carbon footprint are with nuclear energy and more efficient data centers.

This is the second of a two-part mini-series on AI’s environmental footprint. Listen to Part 1 here.

Have a question about AI and the environment? Email us at [email protected] — we’d love to hear from you!

Listen to Short Wave on Spotify, Apple Podcasts and Google Podcasts.

Listen to every episode of Short Wave sponsor-free and support our work at NPR by signing up for Short Wave+ at plus.npr.org/shortwave.

Today’s episode was produced by Hannah Chinn and edited by Rebecca Ramirez. Tyler Jones checked the facts. Kwesi Lee was the audio engineer. Special thanks to Brent Baughman, Julia Simon, Johannes Doerge and the NPR Standards team, as well as to TED Conferences LLC.