Using Generative AI Might Have Serious Environmental Concerns

Generative AI has undeniably transformed various sectors, becoming a ubiquitous force in today’s technological landscape. Among its prevalent applications, image generation stands out as a cornerstone, offering innovative solutions across industries. However, recent research has shed light on a concerning aspect—the substantial environmental toll of generating images with AI. The energy consumption associated with this process is found to be remarkably high, raising urgent questions about the sustainability of such practices. As the world leans heavily on generative AI for creativity and problem-solving, understanding and addressing its environmental impact becomes imperative for responsible and sustainable technological advancement. Here is what this MIT Technology Review article had to emphasize on the issue.

According to the article, each use of AI to generate images carries a significant environmental cost, equivalent to fully charging a smartphone. Researchers from Hugging Face and Carnegie Mellon University found that generating text with AI is less energy-intensive, with 1,000 text generations using only 16% of a smartphone charge. The article highlights that the carbon footprint of AI comes mainly from its usage, not just the energy-intensive training phase. Generating images, especially with powerful AI models, is the most energy-intensive AI task, contributing to significant carbon emissions, according to the article. Finally, the article urges thoughtful generative AI usage, favoring specialized, less carbon-intensive models over larger, more versatile ones to reduce environmental impact.

Balancing the remarkable advancements facilitated by generative AI with a conscientious approach toward energy consumption emerges as a pivotal consideration for a sustainable technological future. Read through the preceding text to get to know more.