Researchers have found that the carbon footprint of generative AI-based tools that can turn text prompts into images and videos is far worse than we previously thought.<\/p>\n
As detailed in a new paper<\/a>, researchers from the open-source AI platform Hugging Face found that the energy demands of text-to-video generators quadruple when the length of a generated video doubles \u2014 indicating that the power required for increasingly sophisticated generations doesn\u2019t scale linearly.<\/p>\n For instance, a six-second AI video clip consumes four times as much energy as a three-second clip.<\/p>\n \u201cThese findings highlight both the structural inefficiency of current video diffusion pipelines and the urgent need for efficiency-oriented design,\u201d the researchers concluded in their paper. <\/p>\n Experts are warning that we\u2019re rolling out generative AI tools without a full grasp of their true environmental impacts.<\/p>\n \u201cUltimately, we found that the common understanding of AI\u2019s energy consumption is full of holes,\u201d MIT Technology Review wrote in a recent analysis<\/a>. <\/p>\n While image generators used the equivalent of five seconds of microwave warming<\/a> to generate a single 1,024 x 1,024 pixel image, video generators proved far more energy-intensive. To spit out a five-second clip, the researchers found that it takes the equivalent of running a microwave for over an hour. If they\u2019re consuming far more power as the length increases, the math doesn\u2019t look good.<\/p>\n Those demands rise even faster for longer clips, implying \u201crapidly increasing hardware and environmental costs,\u201d according to the Hugging Face researchers\u2019 paper.<\/p>\n Fortunately, there are ways to slim down those demands, including intelligent caching, the reusing of existing AI generations, and \u201cpruning<\/a>,\u201d meaning the sifting out of inefficient examples from training datasets.<\/p>\n But whether those efforts will be enough to make a dent in the enormous electricity consumption<\/a> of current AI tools remains to be seen. The scale of its impact is substantial, with AI-related energy usage already representing 20 percent of global datacenter power demands, according to a recent study<\/a>.<\/p>\n Meanwhile, tech giants are investing tens of billions of dollars<\/a> in infrastructure buildouts, sometimes abandoning climate goals in the process. In its 2024 environmental impact report<\/a>, Google admitted<\/a> that it was woefully behind its ambitious plan to reach net-zero carbon emissions by 2030, seeing a staggering 13 percent increase in carbon emissions year over year, in large part due to its embrace of generative AI.<\/p>\n Earlier this year, the company released its Veo 3 AI video generator, later boasting<\/a> that users had created over 40 million videos in just seven weeks.<\/p>\n While the environmental impact of the tool remains unknown \u2014 Google isn\u2019t exactly incentivized to investigate its sizable contributions to carbon emissions \u2014 chances are it\u2019s far worse than we think<\/a>.<\/p>\n