meta_pixel
Tapesearch Logo
Log in
Marketplace Tech

AI’s carbon footprint is growing. Is it worth it?

Marketplace Tech

Marketplace

News, Technology

4.51.3K Ratings

🗓️ 20 April 2023

⏱️ 7 minutes

🧾️ Download transcript

Summary

Between mining for rare minerals, cooling data centers, and running computers for millions of hours, the climate impact of artificial intelligence is big and getting bigger. Marketplace’s Meghan McCarty Carino spoke with Sasha Luccioni, the climate lead for the AI company Hugging Face, about the process of training an earlier version of ChatGPT, which emitted roughly the same amount of carbon dioxide as a gas-powered car driving over one million miles.

Transcript

Click on a timestamp to play from that location

0:00.0

Marketplace Morning Reports new Skin in the Game series explores what we can learn about

0:04.6

money and careers from the $300 billion video game industry. Plus, here how an Oakland-based

0:11.0

program helps young people get the skills they need to break into this booming industry.

0:15.9

Listen to Skin in the Game and more from the Marketplace Morning Report wherever you get your

0:20.7

podcasts. You might want to consider the carbon footprint before you ask a chatbot to write bad

0:29.2

high coups and stuff. From American public media, this is Marketplace Tech. I'm Megan McCarty Carino.

0:44.7

First, there are the resources needed to make the power-hungry chips. These artificial intelligence

0:50.2

tools run on, then hours of intensive computing to train these large language models,

0:56.4

and the cloud storage to keep them connected to millions of users. That all adds up,

1:02.7

and Sasha Lucioni has been doing the map. She's the climate lead at the AI company Hugging Face.

1:09.2

And she says just the process of training an earlier version of chat GPT emitted about the same

1:15.6

CO2 as a gas car driving more than a million miles. Especially new generations of AI models,

1:22.6

they take literally millions of hours of compute time to train. That uses a lot of energy to produce

1:29.3

that energy. You have to use some kind of natural gas or oil or coal or hopefully hydrial

1:36.3

electricity, but that's one big source of emissions and also the hardware that's being used,

1:41.2

so to create these GPU special chips that you need for training AI models, it uses a lot of

1:47.2

rare metals. It uses a lot of water as well. And the same thing goes for cooling GPU servers,

1:53.2

they actually also need to be constantly cooled with water in order to prevent them from overheating.

1:57.5

And so there's all these environmental aspects. But since AI is so ephemeral and it's this

2:04.0

very vague concept, and people have trouble realizing that it comes with physical consequences and

2:09.2

impacts. And I get that because it's like it doesn't have a body, right? But actually it's running

2:14.1

somewhere on a cloud server that's being powered by coal electricity. It's just that we can't see it.

...

Please login to see the full transcript.

Disclaimer: The podcast and artwork embedded on this page are from Marketplace, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of Marketplace and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.

Copyright © Tapesearch 2026.