Kansas: A new study has found that tasks such as writing and imaging performed by artificial intelligence emit hundreds of times less carbon than tasks performed by humans.
A joint study by the University of Kansas and the University of California-Irvine compared artificial intelligence systems ChatGTP, BloomAI, DALL-E2 and others that perform text and image tasks for humans. .
Like cryptocurrencies, artificial intelligence is also under discussion due to its contribution to energy consumption and climate change. Human-caused emissions and environmental impact have long been examined, but comparisons between the two are rare.
However, a comparison by the researchers found that the amount of carbon dioxide emissions per page written by artificial intelligence systems was 130 to 1,500 times lower than that written by humans. While image-forming systems produce 310 to 2900 times less carbon dioxide than humans.
The research, published in the journal Nature, was conducted jointly by Andrew Torrance of the University of Kansas and Bill Tomlinson, Rebecca Black and Donald Patterson of the University of California-Irvine.
