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.
As detailed in a new paper, 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 — indicating that the power required for increasingly sophisticated generations doesn’t scale linearly.
For instance, a six-second AI video clip consumes four times as much energy as a three-second clip.
“These findings highlight both the structural inefficiency of current video diffusion pipelines and the urgent need for efficiency-oriented design,” the researchers concluded in their paper.
Experts are warning that we’re rolling out generative AI tools without a full grasp of their true environmental impacts.
“Ultimately, we found that the common understanding of AI’s energy consumption is full of holes,” MIT Technology Review wrote in a recent analysis.
While image generators used the equivalent of five seconds of microwave warming 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’re consuming far more power as the length increases, the math doesn’t look good.
Those demands rise even faster for longer clips, implying “rapidly increasing hardware and environmental costs,” according to the Hugging Face researchers’ paper.
Fortunately, there are ways to slim down those demands, including intelligent caching, the reusing of existing AI generations, and “pruning,” meaning the sifting out of inefficient examples from training datasets.
But whether those efforts will be enough to make a dent in the enormous electricity consumption 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.
Meanwhile, tech giants are investing tens of billions of dollars in infrastructure buildouts, sometimes abandoning climate goals in the process. In its 2024 environmental impact report, Google admitted 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.
Earlier this year, the company released its Veo 3 AI video generator, later boasting that users had created over 40 million videos in just seven weeks.
While the environmental impact of the tool remains unknown — Google isn’t exactly incentivized to investigate its sizable contributions to carbon emissions — chances are it’s far worse than we think.
More on AI energy usage: How Much Electricity It Actually Takes to Use AI May Surprise You