The report examines claims made by tech companies that generative AI can aid in combating climate change, highlighting a distinction between traditional machine learning and energy-intensive generative models. An analysis of 154 statements found no credible evidence that tools like Google’s Gemini or Microsoft’s Copilot lead to significant reductions in greenhouse gas emissions. Ketan Joshi, an energy analyst, criticized these claims as “greenwashing,” similar to fossil fuel firms promoting minor renewable investments while mainly contributing to pollution.
The study, commissioned by organizations such as Beyond Fossil Fuels, emphasized the misleading portrayal of AI’s climate benefits, with most claims lacking solid evidence. Only 26% of the green claims referenced academic research, and many lacked any citations. Although data centers currently use only 1% of global electricity, their share in the U.S. is expected to rise significantly, raising concerns about their environmental impact.
Experts call for greater transparency regarding the carbon footprint of AI technologies, urging the industry to differentiate between the positive impacts of traditional AI and the negative implications of generative AI. Joshi argued that confusing these distinctions diverts attention from the substantial environmental harm caused by data center expansion.
Source link
