Researchers have constructed a pair of AI agents that can calculate the carbon footprint of an electronic device in no more time and while generating no more emissions than brewing a cup of tea.<\/p>\n
The advance could be a boon for consumers who want to purchase more sustainable cell phones and computers<\/a>. Information about the carbon footprint of these items has been hard to come by until now because electronic devices are usually made of hundreds of components\u2014chips and circuit boards and screens and cases and so on.<\/p>\n The emissions impact of each of these components has to be totted up separately, and sometimes data on components isn\u2019t publicly available or doesn\u2019t exist at all. It can take human life-cycle analysis experts months to assemble the information necessary to calculate the carbon footprint of a single electronic device.<\/p>\n So the researchers built a pair of AI agents\u2014computer programs that solve problems autonomously\u2014to mimic the process that human experts use to create environmental assessments of products. It\u2019s all part of an effort to \u201c[create] a future where individuals can understand the carbon footprint of a product as easily as looking at food nutrition labels, and companies can make informed decisions to create more sustainable products,\u201d says study team member Vikram Iyer<\/a>, a computer scientist at the University of Washington in Seattle.<\/p>\n Iyer and his collaborators constructed one AI agent to act as the project manager, specifying what information is needed and how it will be assembled and analyzed. The other agent is a sort of gofer, searching online for product descriptions, images, and other documents that contain information about a given device and the components it is made of.<\/p>\n All the \u201cteam\u201d needs to get started is a model name or product photo. They complete their work in roughly a minute, the researchers report.<\/p>\n \u00a0<\/p>\n \u00a0<\/p>\n The system cleverly leverages public sources of data that aren\u2019t typically used in life-cycle analyses, such as specs from government agency databases and images of the insides of electronic devices<\/a> from volunteer repair communities such as iFixit and YouTube videos. \u201cThis is where AI really shines by helping automatically sift through this data,\u201d Iyer says.<\/p>\n The resulting estimates of environmental impact are well within the range of variation typically seen in life-cycle analyses conducted by human experts, but emerge at a much faster pace.<\/p>\n Developing this approach also led the researchers to two ways to improve the process even further. First, it turns out that common devices like smartphones and laptops are often made of parts<\/a> from only a small number of companies. So products with similar specifications like screen size and processor type tend to have very similar carbon footprints. The researchers developed a way of quickly estimating the carbon footprint of a device based on a weighted average of similar products.<\/p>\n \u201cThis insight is very helpful to go from zero information to a ballpark estimate for designers early in the product development process, or for consumers trying to find information on a product with no sustainability information,\u201d Iyer says.<\/p>\n Second, they developed a more rigorous way to fill in missing data. If the carbon emissions associated with a given material such as a certain type of plastic used in a laptop\u2019s casing are unknown, human life-cycle analysis experts typically make an estimate based on a similar one. But this can lead to inaccuracies when two materials, say, have a similar name but are actually very different. The new method chooses the best proxy based on physical properties and other characteristics.<\/p>\n As well as helping consumers find the most sustainable products to purchase, the system could free up corporate sustainability teams to spend more time on reducing the carbon footprint of their products rather than painstakingly calculating the footprint of existing ones, the researchers say.<\/p>\n Source: Zhang Z.\u00a0et al.\u00a0\u201cSustainability assessment using multimodal artificial intelligence agents.<\/a>\u201d\u00a0Nature Electronics\u00a02026.<\/p>\n Image: \u00a9 Anthropocene Magazine (based on Photo by Kelly Sikkema<\/a> on Unsplash<\/a>)<\/p>\n","protected":false},"excerpt":{"rendered":"Researchers have constructed a pair of AI agents that can calculate the carbon footprint of an electronic device…\n","protected":false},"author":2,"featured_media":84042,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[405,7537,20778,45342,6931,45343,45344],"class_list":["post-84041","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-ai-agents","tag-artificial-intelligence-agents","tag-carbon-footprint","tag-daily-science","tag-electronics","tag-energy-decarbonization","tag-sustainable-consumption"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/84041","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=84041"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/84041\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/84042"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=84041"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=84041"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=84041"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}