Deep Learning has a Terrible Carbon Footprint

Dr. Joy Buolamwini writes on her LinkedIn post:

I’ve been thinking about the hope expressed that AI can help mitigate climate change. 

Yes, AI can be used to find more efficient ways to use energy. DeepMind proved itself useful to Google in 2016 by reducing data center cooling costs by 40%.

At the same time, we already know that developing some types of AI systems has a substantial environmental impact. 

“ …researchers at the University of Massachusetts, Amherst, performed a life cycle assessment for training several common large AI models. They found that the process can emit more than 626,000 pounds of carbon dioxide equivalent—nearly five times the lifetime emissions of the average American car (and that includes manufacture of the car itself).”

I would like to hear from those who have been exploring the possibilities of using AI to help mitigate climate change. Are there AI offsets in the works? Would they just be a stop gap? What are the most compelling examples you have seen of AI being used to combat climate change?

Please click on this link to check out the post.

Image credit: Image by Freepik