Post by Paul G. Allen School of Computer Science & Engineering

Ever wonder about the environmental impact of that smartphone, laptop or other electronic device you use everyday? According to the United Nations Environment Program, as a whole the electronics industry generates around 41 MILLION TONS of e-waste annually — and only a fraction of that is recycled. For companies and consumers who want to make more sustainable choices, it can be difficult and time-consuming to collect and compare information about the carbon footprint of various devices. But that may be about to change. A team of University of Washington researchers and their collaborators developed a system in which a pair of AI agents work together to assemble what’s known in the industry as life cycle assessments, or LCAs. Instead of taking days (or longer) to collect the necessary data, the AI agents can quickly comb publicly available data from sources such as product descriptions, images of device components and other online documents to come up with an estimate in about a minute. And for those who wonder about the impact of the AI itself, the researchers thought of that, too. The system is designed to run on smaller, less energy-intensive models, and it first checks to see if there is an existing LCA before activating the agents. Once activated, the process of generating a new LCA has about the same carbon footprint as brewing a cup of tea. The team published its paper in the Nature Portfolio journal Nature Electronics. #UWAllen and University of Washington College of Engineering affiliated co-authors include Zhihan Zhang, Vikram Iyer, Alexander Metzger, Zachary Englhardt, Yuxuan Mei, Adriana Schulz, Felix Hahnlein and Shwetak Patel. Read the UW News story: https://lnkd.in/e_Y58Bxy #Research #Sustainability #SustainableComputing #SustainableAI #AgenticAI #AIforGood #UWdiscovers (Photo by Hal Gatewood on Unsplash)