How much does your own AI use matter? With all the warnings about AI’s adverse impact on the environment, it can be tough to understand what that means at the individual level. In this episode, Morgan breaks down the hidden costs of generative AI into something more relatable: microwave time. She’s joined by MIT Technology Review reporters Casey Crownhart and James O’Donnell, who spent months investigating how much energy and water AI systems actually use.

Together, they unpack how AI models are trained and which ones are more resource-intensive, what effect the expansion of AI data centers has on local energy grids and just how much electricity it takes when we ask AI to generate text, images and videos.

Guests:

Casey Crownhart, senior climate reporter at MIT Technology Review

James O'Donnell, senior AI reporter at MIT Technology Review

Further Reading:

We did the math on AI’s energy footprint. Here’s the story you haven’t heard. — Casey Crownhart and James O’Donnell, MIT Technology Review

 AI Energy Score v2: Refreshed Leaderboard, now with Reasoning 🧠 — 

Sasha Luccioni and Boris Gamazaychikov, Hugging Face

Stop worrying about your AI footprint. Look at the big picture instead. — Casey Crownhart, MIT Technology Review 

Google says a typical AI text prompt only uses 5 drops of water — experts say that’s misleading — Justine Calma, The Verge

Read the Transcript here

Want to give us feedback on the show? Shoot us an email at CloseAllTabs@KQED.org

Follow us on ⁠Instagram⁠ and ⁠TikTok⁠

Credits: Close All Tabs is hosted by Morgan Sung. Our team includes producer Maya Cueva, editor Chris Hambrick and senior editor Chris Egusa who also composed our theme song and credits music. Additional music from APM. Audio engineering by Brendan Willard. Audience engagement support from Maha Sanad. Jen Chien is our Director of Podcasts. Katie Sprenger is our Podcast Operations Manager. Ethan Toven-Lindsey is our Editor in Chief.

Learn more about your ad choices. Visit megaphone.fm/adchoices