Carbon emissions of writing and illustrating are lower for AI than for humans \ stacker news ~science

pull down to refresh

Carbon emissions of writing and illustrating are lower for AI than for humans www.nature.com/articles/s41598-024-54271-x?fbclid=IwZXh0bgNhZW0CMTEAAR2BMi8WbArASVW8CPDXi7TN3RviTIh2UOZIJNxVH_tnAzOBbVz6bfrA8SU_aem_DSqp9fcelU4mSSVgx8pU6g

505 sats \ 10k boost \ 13 comments \ @south_korea_ln 19 Nov science

This one is going to be milked by the AI company PR teams ;)

Abstract

As AI systems proliferate, their greenhouse gas emissions are an increasingly important concern for human societies. In this article, we present a comparative analysis of the carbon emissions associated with AI systems (ChatGPT, BLOOM, DALL-E2, Midjourney) and human individuals performing equivalent writing and illustrating tasks. Our findings reveal that AI systems emit between 130 and 1500 times less CO2e per page of text generated compared to human writers, while AI illustration systems emit between 310 and 2900 times less CO2e per image than their human counterparts. Emissions analyses do not account for social impacts such as professional displacement, legality, and rebound effects. In addition, AI is not a substitute for all human tasks. Nevertheless, at present, the use of AI holds the potential to carry out several major activities at much lower emission levels than can humans.

Food for thought...

EDIT: interesting disclaimer :)

view all related items

136 sats \ 7 replies \ @SimpleStacker 19 Nov

For this study, we included the hardware and energy used to provide the AI service, but not the software development cycle or the software engineers and other personnel who worked on the AI. This choice is analogous to how, with the human writer, we included the footprint of that human’s life, but not their parents.

We considered the energy consumption and carbon emissions associated with human activities involved in writing and illustrating tasks. In particular, this assessment included factors such as the annual energy footprint of residents of various regions.

Literally, wtf? You consider the annual energy footprint of humans doing the writing, but not the human energy footprint of AI development? That doesn't seem like an apples to apples comparison.

I'm sorry to say, but I think Nature should be embarrassed to publish this...

37 sats \ 4 replies \ @south_korea_ln OP 19 Nov

It's in Scientific Reports, the pay-to-publish journal of the Nature family. They likely paid ~2500 dollars to get this in. Good incentive for the journal to be less selective on what gets accepted by the Editor for further review. And after that, it just takes 2-3 not too negative reports by the referees to get it published.

5 sats \ 3 replies \ @SimpleStacker 19 Nov

Interesting. I didn't know there was a pay-to-publish arm of Nature. Maybe I should try to get an article in lol

37 sats \ 2 replies \ @south_korea_ln OP 19 Nov

I only found out recently as we were looking to see if a "negative results" journal exists in physics. ChatGPT recommended this one :) We didn't go for it though, so no personal experience with it, and thus can't compare if the process is actually easier than with their more legit Nature or Nature Physics journals.

0 sats \ 1 reply \ @SimpleStacker 19 Nov

I work in a business school so most people don't publish in Nature and aren't familiar with these nuances. I think if I could get one in, people would automatically be super impressed, even if it's "Scientific Reports" lol.

0 sats \ 0 replies \ @south_korea_ln OP 19 Nov

That's part of the appeal they're banking on.

Well now you know what to spend your next salary on ;)

0 sats \ 1 reply \ @SimpleStacker 19 Nov

I should also add...

The parent analogy makes no sense. Human writers would have been born regardless of whether they became writers. But the AI developers wouldn't have worked on development except for the AI.

Furthermore, calculating the carbon footprint of writing as the annual carbon footprint multiplied by time spent writing makes no sense either. Those human writers likely would have been using the same energy during the time spent writing even if they weren't writing.

0 sats \ 0 replies \ @south_korea_ln OP 19 Nov

Thanks for saving me the trouble of actually reading the article. Or maybe, now i wanna read it even more to see this kind of questionable hypotheses with my own eyes.

Reminds me of how media often reports on certain one-time events: "the presence of the police at this football match cost the tax payer 1 million dollars." Well, you would have had to pay these police officers, regardless of if there had been a football match or no match, no?

131 sats \ 1 reply \ @cryotosensei 19 Nov

What is it about this article that prompts you to pay a 10k boost? Haha

0 sats \ 0 replies \ @south_korea_ln OP 19 Nov

Experimenting with boosting an article in my own territory. Wanted to check that indeed, I'm getting most of it back as reward. Will know at midnight, but should be 7k out of the 10k.

136 sats \ 1 reply \ @xz 19 Nov

To extrapolate the notion further, would it not be correct to say that the logical conclusion is an automated planet of artificial intelligence would be less carbon intensive than a planet with human activity, or biological activity.