How Businesses Are Awakening to the Environmental Impact of Generative AI

Businesses are starting to notice the environmental impact of artificial intelligence, particularly the energy and resources needed to power data centers. Recent research highlights how the growing interest in AI has raised concerns about its heavy consumption of electricity and water for cooling.

A study by Capgemini reveals that almost half (48%) of executives believe that their generative AI projects contribute to rising greenhouse gas emissions. However, while awareness is increasing, measuring this impact remains difficult. Only 12% of companies actually track the carbon footprint of their AI efforts. Sustainability often takes a backseat when organizations make decisions about which AI models to adopt.

Many companies depend on third-party vendors, which limits their control over emissions. About 75% of organizations use pre-trained models instead of creating their own. Some businesses are looking at smaller models or renewable energy choices to lessen their environmental effects, yet sustainability isn’t a major priority for most executives. Just one in five considers it an essential factor in decision-making.

Furthermore, many model providers don’t share enough data regarding their environmental impact. The way data sets are bundled makes it challenging to pinpoint specific effects.

To address these issues, Capgemini suggests that companies start focusing on smaller, task-oriented AI models to decrease energy use. They can also enhance sustainability by selecting more efficient hardware and opting for green data centers.

The report emphasizes the need for clearer guidelines on the ethical and sustainable use of generative AI. Cyril Garcia, Capgemini’s Head of Global Sustainability Services, states, “For generative AI to add real value in a sustainable way, we need discussions on data collaboration and industry standards. This will help business leaders make informed choices and reduce their environmental impact.”

Source link