AI Image Generation Carries High Environmental Cost, Study Finds
Research reveals that creating a single AI image consumes as much energy as charging a smartphone, with popular models potentially producing more carbon than a short car journey.
- A new study reveals that generating a single AI image consumes as much energy as charging a smartphone, leading to a significant carbon footprint.
- The research, conducted by Hugging Face and Carnegie Mellon University, examined 13 AI tasks and found that image generation was the most energy-intensive.
- Popular AI models like ChatGPT’s Dall-E and Midjourney could produce more carbon than driving 4.1 miles.
- The study urges AI practitioners to be transparent about the environmental impacts of their models.
- Despite the environmental costs, AI could also help tackle environmental challenges by monitoring emissions and tracking air quality.