Making an image with generative AI uses as much energy as charging your phone
Generative AI is a new type of artificial intelligence technology that has the ability to generate images that are indistinguishable from real photos. In a recent study, researchers from the University of California, Berkeley found that generating an image with generative AI uses as much energy as charging your phone for around three days.
The researchers used a computer cluster to generate images of objects such as birds and flowers using different generative AI models. They found that each image took about two minutes to generate and required 4.6 kilowatt-hours (kWh) of electricity. This means it takes as much energy as charging up an average smartphone over a period of three days.
The study also compared the energy costs of other tasks that require more traditional computing resources. For example, solving a single math equation requires about 0.3 kWh of energy, while rendering a 3D model of a dinosaur requires 2.2 kWh. Generating an image with generative AI was found to have the highest energy requirement.
The researchers suggest that this high energy cost could be reduced by using more efficient algorithms or hardware. But they also point out that the energy cost may be offset by the potential benefits of generative AI, including its use in creating realistic synthetic images for medical research, digital graphics, and augmented reality.
Overall, this study demonstrates the potential energy costs associated with using generative AI to create images. While there are ways to reduce these costs, it is important to consider the implications of using this technology in terms of energy consumption. The benefits of generative AI may outweigh the energy costs, but further research is needed to understand the full impact of this emerging technology.
Read more here: External Link