{"id":32350,"date":"2026-01-21T17:01:45","date_gmt":"2026-01-21T17:01:45","guid":{"rendered":"https:\/\/lamarr-institute.org\/publication\/ground-truthing-ai-energy-consumption-validating-codecarbon-against-external-measurements\/"},"modified":"2026-09-17T11:20:57","modified_gmt":"2026-09-17T11:20:57","slug":"ground-truthing-ai-energy-consumption-validating-codecarbon-against-external-measurements","status":"publish","type":"publication","link":"https:\/\/lamarr-institute.org\/de\/publication\/ground-truthing-ai-energy-consumption-validating-codecarbon-against-external-measurements\/","title":{"rendered":"Ground-Truthing {AI} Energy Consumption: Validating {CodeCarbon} Against External Measurements"},"content":{"rendered":"<p>Machine learning (ML) and artificial intelligence (AI) technology offers opportunities for innovation, however its rapid adoption also significantly impacts our environment. In response to growing resource-awareness, tools such as CodeCarbon were developed to estimate the energy consumption and carbon emissions of AI. While easy to incorporate into research, these tools however make pragmatic assumptions and neglect overhead factors, raising the question of estimation accuracy. This study systematically evaluates the reliability of static and dynamic energy estimation approaches through comparisons with ground-truth measurements across hundreds of AI experiments. By comparing existing tools against affordable energy meters and smart sockets, investigative insights into AI resource demands and estimation inaccuracies are provided. While generally following the general energy consumption trends, well-established tools are shown to make estimation errors of up to 40\\%. By providing empirical evidence on the hidden figures in environmental impact assessments and equipping practitioners with guidelines and code for their experiments, this study promotes transparency and awareness for sustainable AI development.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Machine learning (ML) and artificial intelligence (AI) technology offers opportunities for innovation, however its rapid adoption also significantly impacts our environment. In response to growing resource-awareness, tools such as CodeCarbon were developed to estimate the energy consumption and carbon emissions of AI. While easy to incorporate into research, these tools however make pragmatic assumptions and neglect overhead factors, raising the question of estimation accuracy. This study systematically evaluates the reliability [&hellip;]<\/p>\n","protected":false},"author":14,"featured_media":0,"template":"","meta":{"_acf_changed":false,"footnotes":""},"publication-type":[30],"class_list":["post-32350","publication","type-publication","status-publish","hentry","publication-type-article"],"acf":[],"publishpress_future_workflow_manual_trigger":{"enabledWorkflows":[]},"_links":{"self":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication\/32350","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication"}],"about":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/types\/publication"}],"author":[{"embeddable":true,"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/users\/14"}],"version-history":[{"count":0,"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication\/32350\/revisions"}],"wp:attachment":[{"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/media?parent=32350"}],"wp:term":[{"taxonomy":"publication-type","embeddable":true,"href":"https:\/\/lamarr-institute.org\/de\/wp-json\/wp\/v2\/publication-type?post=32350"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}