{"id":29914,"date":"2025-06-19T04:00:03","date_gmt":"2025-06-19T04:00:03","guid":{"rendered":"https:\/\/medexperts.pro\/?p=29914"},"modified":"2025-06-19T04:25:22","modified_gmt":"2025-06-19T04:25:22","slug":"your-a-i-queries-come-with-a-climate-cost","status":"publish","type":"post","link":"https:\/\/medexperts.pro\/?p=29914","title":{"rendered":"Your A.I. Queries Come With a Climate Cost"},"content":{"rendered":"
<\/div>\n

When it comes to artificial intelligence, more intensive computing uses more energy, producing more greenhouse gases. <\/p>\n

\n
\n
\n

From uninvited results at the top of your search engine queries to offering to write your emails and helping students do homework, generative A.I. is quickly becoming part of daily life as tech giants race to develop the most advanced models and attract users.<\/p>\n

All those prompts come with an environmental cost: A report last year from the Energy Department found A.I. could help increase the portion of the nation\u2019s electricity supply consumed by data centers from 4.4 percent to 12 percent by 2028. To meet this demand, some power plants are expected to burn more coal and natural gas<\/a>.<\/p>\n

And some chatbots are linked to more greenhouse gas emissions than others. <\/span>A study published Thursday in the journal Frontiers in Communication analyzed different generative A.I. chatbots\u2019 capabilities and the planet-warming emissions generated from running them. Researchers found that chatbots with bigger \u201cbrains\u201d used exponentially more energy and also answered questions more accurately \u2014 up until a point.<\/p>\n

\u201cWe don\u2019t always need the biggest, most heavily trained model, to answer simple questions. Smaller models are also capable of doing specific things well,\u201d said Maximilian Dauner, a Ph.D. student at the Munich University of Applied Sciences and lead author of the paper. \u201cThe goal should be to pick the right model for the right task.\u201d<\/p>\n<\/div>\n<\/div>\n

<\/div>\n
\n
\n

The study evaluated 14 large language models, a common form of generative A.I. often referred to by the acronym LLMs, by asking each a set of 500 multiple choice and 500 free response questions across five different subjects. Mr. Dauner then measured the energy used to run each model and converted the results into carbon dioxide equivalents based on global averages.In most of the models tested, questions in logic-based subjects, like abstract algebra, produced the longest answers \u2014 which likely means they used more energy to generate compared with fact-based subjects, like history, Mr. Dauner said.<\/p>\n

A.I. chatbots that show their step-by-step reasoning while responding tend to use far more energy per question than chatbots that don\u2019t. The five reasoning models tested in the study did not answer questions much more accurately than the nine other studied models. The model that emitted the most, DeepSeek-R1, offered answers of comparable accuracy to those that generated a fourth of the amount of emissions.<\/p>\n<\/div>\n<\/div>\n

\n
\n
\n

<\/p>\n

\n

\t<!–><\/p>\n

[–><\/p>\n
\n
<\/p>\n
<\/p>\n

<\/h3>\n

<\/div>\n

<\/div>\n<\/div>\n

<\/p>\n

\n
<\/p>\n
<\/p>\n
<!–><\/p>\n
[!–><\/div>\n

<\/div>\n<\/div>\n

<!–><\/div>\n<\/div>\n

–> <\/p>\n

\n
<\/p>\n
<\/p>\n

<\/p>\n

<\/p>\n

<\/p>\n

<\/p>\n

<\/p>\n

<\/div>\n

<\/div>\n<\/div>\n

<!–><\/figure>\n

–> <\/p><\/div>\n<\/div>\n<\/div>\n<\/section>\n<\/div>\n

\n
\n
\n

<\/p>\n

\n

\t<!–><\/p>\n

[–><\/p>\n
\n
<\/p>\n
<\/p>\n

<\/h3>\n

<\/p>\n

<\/p>\n

<\/div>\n

<\/div>\n<\/div>\n

<\/p>\n

\n
<\/p>\n
<\/p>\n
<!–><\/p>\n
\n
\n\n\n\n<\/p>\n\n
<\/th>\n

[–><\/p>\n

Abstract algebra<\/th>\nH.S. math<\/th>\nH.S. world history<\/th>\nInt\u2019l law<\/th>\nPhilosophy<\/th>\n

<\/p>\n

\n
Average of 14 language models<\/div>\n

<\/td>\n

<\/p>\n

\n
1.0<\/span><\/div>\n<\/td>\n
\n
1.0<\/span><\/div>\n<\/td>\n
\n
0.6<\/span><\/div>\n<\/td>\n
\n
0.5<\/span><\/div>\n<\/td>\n
\n
0.4<\/span><\/div>\n<\/td>\n

<\/tr>\n

<\/tbody>\n

<\/div>\n

<!–> <\/p>\n

\n
Results by model<\/div>\n<\/div>\n
\n\n\n\n<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
<\/th>\n

[–><\/p>\n

<\/th>\n<\/th>\n<\/th>\n<\/th>\n<\/th>\n

<\/p>\n

\n
Deepseek-r1 70B<\/div>\n

<\/p>\n

Reasoning<\/div>\n

<\/td>\n

<\/p>\n

\n
4.8<\/span><\/div>\n<\/td>\n
\n
4.0<\/span><\/div>\n<\/td>\n
\n
1.7<\/span><\/div>\n<\/td>\n
\n
1.4<\/span><\/div>\n<\/td>\n
\n
1.2<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Cogito 70B<\/div>\n

<\/p>\n

Reasoning<\/div>\n

<\/td>\n

<\/p>\n

\n
2.3<\/span><\/div>\n<\/td>\n
\n
2.3<\/span><\/div>\n<\/td>\n
\n
1.7<\/span><\/div>\n<\/td>\n
\n
1.6<\/span><\/div>\n<\/td>\n
\n
1.4<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Qwen2.5 72B<\/div>\n

<\/td>\n

<\/p>\n

\n
1.1<\/span><\/div>\n<\/td>\n
\n
0.8<\/span><\/div>\n<\/td>\n
\n
1.0<\/span><\/div>\n<\/td>\n
\n
0.7<\/span><\/div>\n<\/td>\n
\n
0.5<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Cogito 8B<\/div>\n

<\/p>\n

Reasoning<\/div>\n

<\/td>\n

<\/p>\n

\n
1.1<\/span><\/div>\n<\/td>\n
\n
1.8<\/span><\/div>\n<\/td>\n
\n
0.3<\/span><\/div>\n<\/td>\n
\n
0.3<\/span><\/div>\n<\/td>\n
\n
0.3<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Llama 3.3 70B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.7<\/span><\/div>\n<\/td>\n
\n
0.7<\/span><\/div>\n<\/td>\n
\n
0.7<\/span><\/div>\n<\/td>\n
\n
0.6<\/span><\/div>\n<\/td>\n
\n
0.4<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Llama 3.1 70B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.6<\/span><\/div>\n<\/td>\n
\n
0.6<\/span><\/div>\n<\/td>\n
\n
0.7<\/span><\/div>\n<\/td>\n
\n
0.6<\/span><\/div>\n<\/td>\n
\n
0.3<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Deepseek-r1 7B<\/div>\n

<\/p>\n

Reasoning<\/div>\n

<\/td>\n

<\/p>\n

\n
1.1<\/span><\/div>\n<\/td>\n
\n
1.0<\/span><\/div>\n<\/td>\n
\n
0.4<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Deepseek-r1 8B<\/div>\n

<\/p>\n

Reasoning<\/div>\n

<\/td>\n

<\/p>\n

\n
1.1<\/span><\/div>\n<\/td>\n
\n
0.9<\/span><\/div>\n<\/td>\n
\n
0.4<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Cogito 70B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.6<\/span><\/div>\n<\/td>\n
\n
0.5<\/span><\/div>\n<\/td>\n
\n
0.7<\/span><\/div>\n<\/td>\n
\n
0.5<\/span><\/div>\n<\/td>\n
\n
0.4<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Qwen 72B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.3<\/span><\/div>\n<\/td>\n
\n
0.5<\/span><\/div>\n<\/td>\n
\n
0.5<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Llama 3.1 8B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Qwen2.5 7B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.2<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Cogito 8B<\/div>\n

<\/td>\n

<\/p>\n

\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.2<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n

<\/tr>\n

\n
Qwen 7B<\/div>\n

<\/td>\n

<\/p>\n

\n
<0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
0.1<\/span><\/div>\n<\/td>\n
\n
<0.1<\/span><\/div>\n<\/td>\n
\n
<0.1<\/span><\/div>\n<\/td>\n

<\/tr>\n

<\/tbody>\n

<\/p>\n

<\/div>\n

<\/div>\n

<!–><\/p><\/div>\n

–><\/div>\n<\/div>\n

<!–><\/div>\n<\/div>\n

–> <\/p>\n

\n
<\/p>\n
<\/p>\n

<\/p>\n

<\/p>\n

<\/p>\n

<\/p>\n

<\/p>\n

<\/div>\n

<\/div>\n<\/div>\n

<!–><\/figure>\n

–> <\/p><\/div>\n<\/div>\n<\/section>\n<\/div>\n

\n
\n
\n
<\/div>\n
\n
\n
\n

We are having trouble retrieving the article content.<\/p>\n

Please enable JavaScript in your browser settings.<\/p>\n<\/div>\n<\/div>\n

\n
\n

Thank you for your patience while we verify access. If you are in Reader mode please exit and\u00a0log into<\/a>\u00a0your Times account, or\u00a0subscribe<\/a>\u00a0for all of The Times.<\/p>\n

\n<\/div>\n
\n
\n

Thank you for your patience while we verify access.<\/p>\n

Already a subscriber?\u00a0Log in<\/a>.<\/p>\n

Want all of The Times?\u00a0Subscribe<\/a>.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/section>\n","protected":false},"excerpt":{"rendered":"

When it comes to artificial intelligence, more intensive computing uses more energy, producing more greenhouse gases. From uninvited results at the top of your search engine queries to offering to write your emails and helping students do homework, generative A.I. is quickly becoming part of daily life as tech giants race to develop the most advanced models and attract users.All those prompts come with an environmental cost: A report last year from the Energy Department found A.I. could help increase the portion of the nation\u2019s electricity supply consumed by data centers from 4.4 percent to 12 percent by 2028. To meet this demand, some power plants are expected to burn more coal and natural gas.And some chatbots are linked to more greenhouse gas emissions than others. A study published Thursday in the journal Frontiers in Communication analyzed different generative A.I. chatbots\u2019 capabilities and the planet-warming emissions generated from running them. Researchers found that chatbots with bigger \u201cbrains\u201d used exponentially more energy and also answered questions more accurately \u2014 up until a point.\u201cWe don\u2019t always need the biggest, most heavily trained model, to answer simple questions. Smaller models are also capable of doing specific things well,\u201d said Maximilian Dauner, a Ph.D. student at the Munich University of Applied Sciences and lead author of the paper. \u201cThe goal should be to pick the right model for the right task.\u201dThe study evaluated 14 large language models, a common form of generative A.I. often referred to by the acronym LLMs, by asking each a set of 500 multiple choice and 500 free response questions across five different subjects. Mr. Dauner then measured the energy used to run each model and converted the results into carbon dioxide equivalents based on global averages.In most of the models tested, questions in logic-based subjects, like abstract algebra, produced the longest answers \u2014 which likely means they used more energy to generate compared with fact-based subjects, like history, Mr. Dauner said.A.I. chatbots that show their step-by-step reasoning while responding tend to use far more energy per question than chatbots that don\u2019t. The five reasoning models tested in the study did not answer questions much more accurately than the nine other studied models. The model that emitted the most, DeepSeek-R1, offered answers of comparable accuracy to those that generated a fourth of the amount of emissions.<\/p>\n","protected":false},"author":1,"featured_media":29916,"comment_status":"close","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[],"class_list":["post-29914","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology"],"_links":{"self":[{"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/posts\/29914","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/medexperts.pro\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=29914"}],"version-history":[{"count":2,"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/posts\/29914\/revisions"}],"predecessor-version":[{"id":29917,"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/posts\/29914\/revisions\/29917"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/medexperts.pro\/index.php?rest_route=\/wp\/v2\/media\/29916"}],"wp:attachment":[{"href":"https:\/\/medexperts.pro\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=29914"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/medexperts.pro\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=29914"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/medexperts.pro\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=29914"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}