Tech companies are making bold claims about artificial intelligence (AI) helping to prevent climate change. A recent report, however, raises doubts about these assertions. It points out that many AI tools, especially the energy-hungry generative models, aren’t delivering on their promises of reducing emissions.
The report analyzed statements from 154 tech companies. Most of their claims about AI’s environmental benefits are based on machine learning, not the resource-intensive chatbots and image generators that dominate the industry today. Ketan Joshi, an energy analyst behind the analysis, says these claims are misleading, akin to fossil fuel companies overstating their investments in solar energy while continuing to pollute heavily.
Sasha Luccioni, an AI and climate expert at Hugging Face, noted that discussions often group very different types of AI together. She emphasized that while generative AI and large language models can harm the planet, traditional AI tools focused on prediction and analysis can be beneficial.
In the tech industry, evidence supporting these green claims is often weak or unverified. A staggering 36% of such claims lack any reference or evidence. For instance, a frequently cited statistic is that AI could help reduce 5-10% of global greenhouse gas emissions by 2030. This figure, touted by Google, originates from a blog post referencing “experience with clients,” not solid research.
Data centers, which support these AI tools, consume only about 1% of the world’s electricity. However, this number is projected to double in the US by 2035, raising concerns among energy researchers. While using a large language model like ChatGPT for a simple text query is energy-efficient—comparable to running a lightbulb for a minute—more complex tasks consume significantly more energy.
In response to the report, a Google spokesperson asserted that their emissions reduction figures are grounded in robust science, but Microsoft did not comment. Joshi cautioned that it’s time to reconsider the narrative around AI’s supposed climate benefits, as it can divert attention from the serious energy demands of expanding data facilities.
With the rapid development of AI technologies, we need clearer understanding and honest assessments of their environmental impact. As consumers and industries look to the future, transparency will be crucial in ensuring that claims about AI’s role in climate solutions are grounded in reality, not just marketing speak.
For more in-depth insights, check out related research from the International Energy Agency (IEA) here.
