{"id":191570,"date":"2024-12-09T16:22:41","date_gmt":"2024-12-09T16:22:41","guid":{"rendered":"https:\/\/news.talkwithrattan.com\/index.php\/2024\/12\/09\/generative-ai-is-an-energy-hog-is-the-tech-worth-the-environmental-cost\/"},"modified":"2024-12-09T16:22:41","modified_gmt":"2024-12-09T16:22:41","slug":"generative-ai-is-an-energy-hog-is-the-tech-worth-the-environmental-cost","status":"publish","type":"post","link":"https:\/\/news.talkwithrattan.com\/index.php\/2024\/12\/09\/generative-ai-is-an-energy-hog-is-the-tech-worth-the-environmental-cost\/","title":{"rendered":"Generative AI is an energy hog. Is the tech worth the environmental cost?"},"content":{"rendered":"<div style=\"text-align:center\"><img decoding=\"async\" src=\"https:\/\/i0.wp.com\/www.sciencenews.org\/wp-content\/uploads\/2024\/12\/121424_YE-ai-energy_feat.jpg?fit=800%2C450&amp;ssl=1\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"Generative AI is an energy hog. Is the tech worth the environmental cost?\" title=\"Generative AI is an energy hog. Is the tech worth the environmental cost?\" \/><\/div> \r\n<br><br><div data-component=\"video-embed\">\n\t\t\t\t\n\n\n\n\n<p>It might seem like magic. Type a request into ChatGPT, click a button and \u2014 presto! \u2014 here\u2019s a five-paragraph analysis of Shakespeare\u2019s <em>Hamlet<\/em> and, as an added bonus, it\u2019s written in iambic pentameter. Or tell DALL-E about the chimeric animal from your dream, and out comes an image of a gecko-wolf-starfish hybrid. If you\u2019re feeling down, call up <a href=\"https:\/\/www.sciencenews.org\/article\/artificial-intelligence-digital-ghost-ai\">the digital \u201cghost\u201d of your deceased grandmother<\/a> and receive some comfort (<em>SN: 6\/15\/24, p. 10<\/em>).<\/p>\n\n\n\n<p>Despite how it may appear, none of this materializes out of thin air. Every interaction with a chatbot or other generative AI system funnels through wires and cables to a data center \u2014 a warehouse full of server stacks that pass these prompts through the billions (and potentially trillions) of parameters that dictate how a generative model responds.<\/p>\n\n\n<aside class=\"sn-conversion rich-text rich-text--with-sidebar\">\n<style><![CDATA[\n.email-conversion {\n  border: 1px solid #ffcccb;\n  color: white;\n  margin-top: 50px;\n  background-image: url(\"\/wp-content\/themes\/sciencenews\/client\/src\/images\/cta-module@2x.jpg\");\n  padding: 20px;\n  clear: both;\n}\n\n]]><\/style>\n\n\n\n<div class=\"rich-text embedded-conversion-content is-layout-flow wp-block-group-is-layout-flow\"><div class=\"wp-block-group__inner-container\">\n<style><![CDATA[\n#dynamic-wrapper {\n  border: 1px solid #ffcccb;\n  background-image: url(\"https:\/\/www.sciencenews.org\/wp-content\/uploads\/2024\/11\/cta_background_aurora.jpg\");\n  background-size: cover;\n  background-position: center center;\n  padding: 20px;\n  clear: both;\n}\n\n#dynamic-conversion {\n  padding: 20px;\n  background:rgba(0,0,0, 0.5);\n  color: white;\n}\n\n#dynamic-conversion h2 {\n  color: white;\n}\n\np.has-text-align-center a {\n  color: white !important;\n  text-decoration: none;\n  font-weight: bold;\n}\n\n]]><\/style>\n\n\n\n<div id=\"dynamic-wrapper\" class=\"wp-block-group\"><div class=\"wp-block-group__inner-container is-layout-constrained wp-block-group-is-layout-constrained\">\n<div id=\"dynamic-conversion\" class=\"wp-block-group\"><div class=\"wp-block-group__inner-container is-layout-flow wp-block-group-is-layout-flow\">\n\n\n\n\n<h2 class=\"wp-block-heading has-text-align-center\"><strong>Have feedback for Science News?<\/strong><\/h2>\n\n\n\n\n\n\n\n<p class=\"has-text-align-center\">Help us improve by telling us about your experience<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n<\/div><\/div>\n\n\n\n\n<\/div><\/div>\n<\/div><\/div>\n\n\n\n<\/aside>\n\n\n<p>Processing and answering prompts eats up electricity, as does the supporting infrastructure like fans and air conditioning that cool the whirring servers. In addition to big utility bills, the result is a lot of climate-warming carbon emissions. Electricity generation and server cooling also suck up tons of water, which is used in fossil fuel and nuclear energy production, and for evaporative or liquid heat dissipation systems.<\/p>\n\n\n\n<p>This year, as the popularity of generative AI continued to surge, environmentalists sounded the alarm about this resource-hungry technology. The debate over how to weigh the costs against the less tangible benefits that generative AI brings, such as increased productivity and information access, is steeped in ideological divisions over the purpose and value of technology.<\/p>\n\n\n\n<p>Advocates argue this latest revolution in AI is a societal good, even a necessity, that\u2019s bringing us closer than ever to artificial general intelligence, hypercapable computer systems that some argue could be a paradigm-shifting technology on par with the printing press or the internet.<\/p>\n\n\n\n<p>Generative AI \u201cis an accelerator for anything you want to do,\u201d says Rick Stevens, an associate lab director at Argonne National Laboratory and a computer scientist at the University of Chicago. In his view, the tech has already enabled major productivity gains for businesses and researchers.<\/p>\n\n\n\n<p>One analysis found 40 percent gains in performance when skilled workers used AI tools, he notes. AI assistants can boost vocabulary learning in schools, he adds. Or help physicians diagnose and treat patients, and improve access to medical information, says Charlotte Blease, an interdisciplinary researcher at Uppsala University in Sweden who studies health data. Generative AI might even help city planners cut down on traffic (and reduce carbon emissions in the process), or help government agencies better forecast the weather, says Priya Donti, an electrical engineer and computer scientist at MIT and cofounder of the nonprofit Climate Change AI. The list goes on.<\/p>\n\n\n\n<p>Now, at this critical juncture, experts from fields as varied as economics, computer engineering and sustainability are working to assess the true burden of the technology.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">How much energy does AI consume?<\/h2>\n\n\n\n<p>ChatGPT and other generative tools are power hungry, says Alex de Vries, founder of the research and consulting agency Digiconomist and a Ph.D. candidate at Vrije Universiteit Amsterdam. \u201cThe larger you make these models \u2014 the more parameters, the more data \u2014 the better they perform. But of course, bigger also requires more computational resources to train and run them, requiring more power,\u201d says de Vries, who studies the environmental impact of technologies like cryptocurrency and AI. \u201cBigger is better works for generative AI, but it doesn\u2019t work for the environment.\u201d<\/p>\n\n\n\n<p>Training generative AI models to spit out an analysis of Shakespeare or the image of a fantastical animal is costly. The process involves developing an AI architecture, amassing and storing reams of digital data and then having the AI system ingest and incorporate that data \u2014 which can amount to everything publicly available on the internet \u2014 into its decision- making processes. Honing models to be more humanlike <a href=\"https:\/\/www.sciencenews.org\/article\/generative-ai-chatbots-chatgpt-safety-concerns\">and avoid unsafe responses<\/a> takes additional effort (<em>SN: 1\/27\/24, p. 18<\/em>).<\/p>\n\n\n\n<p>All told, training a single model uses more energy than 100 U.S. homes in a year. Querying ChatGPT uses about <a href=\"https:\/\/iea.blob.core.windows.net\/assets\/6b2fd954-2017-408e-bf08-952fdd62118a\/Electricity2024-Analysisandforecastto2026.pdf\" target=\"_blank\" rel=\"noopener\">10 times as much energy as a standard online search<\/a>, according to the International Energy Agency. Composing an email with an AI chatbot can take seven times as much energy as fully charging an iPhone 16, some researchers estimate.<\/p>\n\n\n<aside class=\"sn-conversion rich-text rich-text--with-sidebar\">\n<p class=\"has-text-align-center wp-elements-27c40654034fbeecef6418d6adfe0794\" style=\"color:gray; margin-bottom:0px; font-size:.9rem;\">Sponsor Message<\/p>\n<!-- Tag ID: sciencenews-org_leaderboard_incontent -->\n\n<\/aside>\n\n\n<p>Though training is clearly a big resource suck, when millions of people rely on chatbots for everyday tasks, it adds up, says Shaolei Ren, an electrical and computer engineer at the University of California, Riverside. So much so that the AI sector could soon <a href=\"https:\/\/www.cell.com\/joule\/fulltext\/S2542-4351(23)00365-3\" target=\"_blank\" rel=\"noopener\">draw as much energy annually as the Netherlands<\/a>, de Vries estimated in 2023 in <em>Joule<\/em>. Given generative AI\u2019s rapid growth, the current trajectory already exceeds the prediction.<\/p>\n\n\n\n<div class=\"wp-block-sciencenews-content-sidebar\">\n<h2 class=\"wp-block-heading\">Energy hog<\/h2>\n\n\n\n<p>Answering a single ChatGPT query requires more electricity than a single Google search. <\/p>\n\n\n<iframe loading=\"lazy\" title=\"Energy demands of Google vs. ChatGPT\" class=\"sn-responsive-iframe\" id=\"sn-responsive-iframe-47\" src=\"https:\/\/flo.uri.sh\/visualisation\/20699580\/embed\" width=\"100%\" height=\"300\" layout=\"responsive\" frameborder=\"0\" allowfullscreen=\"\">\n\t<\/iframe>\n\n\n<p>Google search\u2019s total daily energy requirements currently surpass ChatGPT\u2019s because it handles an estimated 8.5 billion searches daily compared with ChatGPT\u2019s 13 million daily queries. Training a generative AI model like ChatGPT is a huge energy suck, but individual usage over time adds up.<\/p>\n\n\n<iframe loading=\"lazy\" title=\"Energy demands of Google vs. ChatGPT\" class=\"sn-responsive-iframe\" id=\"sn-responsive-iframe-48\" src=\"https:\/\/flo.uri.sh\/visualisation\/20699690\/embed\" width=\"100%\" height=\"400\" layout=\"responsive\" frameborder=\"0\" allowfullscreen=\"\">\n\t<\/iframe><\/div>\n\n\n\n<p>And that\u2019s just electricity. <a href=\"https:\/\/arxiv.org\/pdf\/2304.03271\" target=\"_blank\" rel=\"noopener\">Ten to 50 ChatGPT queries use half a liter of water<\/a>, per a 2023 analysis by Ren and colleagues. That turned out to be a big underestimate too, he says, off by a factor of four.<\/p>\n\n\n\n<p>Some engineers and AI experts dispute these numbers. \u201cI don\u2019t understand what the science is behind these [estimates],\u201d says David Patterson, an engineer at Google and professor emeritus at the University of California, Berkeley. \u201cThe only way I can imagine getting an [accurate] answer would be with close cooperation with a company like Google.\u201d<\/p>\n\n\n\n<p>Right now, that\u2019s impossible. Tech companies release limited information about their data centers and AI models, say de Vries and Ren. So it\u2019s hard to precisely assess the cradle-to-grave cost of AI or predict the future. In their estimates, both researchers relied on proxies, such as AI server production numbers from the tech company Nvidia or combining knowledge on data center locations with info from corporate sustainability reports.<\/p>\n\n\n\n<p>Real-world trends, however, do point to AI\u2019s voracious energy appetite. For decades before the generative AI boom, efficiency gains have compensated for the growing energy demand that\u2019s come with expansions in data centers and computing, says Andrew Chien, a computer scientist at the University of Chicago. That\u2019s changed. By the end of 2020, data center expansion began to outpace efficiency improvements, he says. Both Google\u2019s and Microsoft\u2019s self-reported energy usage more than doubled between 2019 and 2023. ChatGPT\u2019s release at the end of 2022 kick-started a generative AI frenzy \u2014 exacerbating the issue, Chien says. Before 2022, total energy demand in the United States had been stable for about 15 years. Now it\u2019s rising.<\/p>\n\n\n\n<p>\u201cThe easiest way to save energy is to not do anything,\u201d Patterson says. But \u201cprogress involves investment and costs.\u201d Generative AI is a very young technology, and stopping now would stymie its potential, he argues. \u201cIt\u2019s too early to know that [generative AI] won\u2019t more than compensate the investment.\u201d<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">A more sustainable path for AI<\/h2>\n\n\n\n<p>The decision need not be between shutting down generative AI development entirely or allowing it to continue unrestrained. Instead, most experts note there\u2019s a more responsible way to approach the technology, mitigating the risks and maximizing the rewards.<\/p>\n\n\n\n<p>Policies requiring companies to disclose where and how they\u2019re using generative AI, as well as the corresponding energy consumption, would be a step in the right direction, says Lynn Kaack, a computer science and public policy expert at the Hertie School in Berlin. Regulating uses of the technology and access to it may prove difficult, but Kaack says that\u2019s key to minimizing environmental and social harm.<\/p>\n\n\n\n<p>Perhaps not everyone, for instance, should be able to freely produce voice clones and photorealistic images with a single click. Should we pour the same amount of resources into supporting a generative meme machine as we do for running a hurricane forecasting model?<\/p>\n\n\n\n<p>More research into the tech\u2019s limitations could also save lots of futile consumption. AI \u201cis very powerful in certain kinds of applications, but completely useless in others,\u201d Kaack says.<\/p>\n\n\n\n<p>Meanwhile, data centers and AI developers could take steps to lessen their carbon emissions and resource use, Chien says. Simple changes like training models only when there\u2019s ample carbon-free power on the grid (say, on sunny days when solar panels produce an excess of energy) or subtly reducing system performance at times of peak energy demand might make a measurable difference. Replacing water-intensive evaporative cooling with liquid- immersion cooling or other closed-loop strategies that allow for water recycling would also minimize demand.<\/p>\n\n\n\n<p>Each of these choices involves trade-offs. More carbon-efficient systems generally use more water, Ren says. There is no one-size-fits-all solution. The alternative to exploring and incentivizing these options \u2014 even if they make it marginally harder for companies to develop ever-bigger AI models \u2014 is risking part of our collective environmental fate, he says.<\/p>\n\n\n\n<p>\u201cThere\u2019s no reason to believe that technology is going to save us,\u201d Chien says \u2014 so why not hedge our bets?<\/p>\n\n\n\n\t\t\t<\/div>\r\n<br>\r\n<br><a href=\"https:\/\/www.sciencenews.org\/article\/generative-ai-energy-environmental-cost\">Source link <\/a>","protected":false},"excerpt":{"rendered":"<p>It might seem like magic. Type a request into ChatGPT, click a button and \u2014 presto! \u2014 here\u2019s a five-paragraph analysis of Shakespeare\u2019s Hamlet and, as an added bonus, it\u2019s written in iambic pentameter. Or tell DALL-E about the chimeric animal from your dream, and out comes an image of a gecko-wolf-starfish hybrid. If you\u2019re [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":191571,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"tdm_status":"","tdm_grid_status":"","fifu_image_url":"https:\/\/i0.wp.com\/www.sciencenews.org\/wp-content\/uploads\/2024\/12\/121424_YE-ai-energy_feat.jpg?fit=800%2C450&ssl=1","fifu_image_alt":"","footnotes":""},"categories":[606],"tags":[5996,2988,3648,1073,62117,2416,1842],"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/posts\/191570"}],"collection":[{"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/comments?post=191570"}],"version-history":[{"count":1,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/posts\/191570\/revisions"}],"predecessor-version":[{"id":191572,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/posts\/191570\/revisions\/191572"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/media\/191571"}],"wp:attachment":[{"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/media?parent=191570"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/categories?post=191570"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/news.talkwithrattan.com\/index.php\/wp-json\/wp\/v2\/tags?post=191570"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}