Tech’s Unproven Claims: How AI Companies Exaggerate Climate Benefits

A new report reveals how major tech companies are making bold, often unsupported claims about artificial intelligence’s potential to combat climate change, while simultaneously driving up emissions through massive data center expansion.

The Problem with Big Tech’s Climate Claims

Energy researcher Ketan Joshi investigated Google’s widely-cited claim that AI could reduce global greenhouse gas emissions by 5-10% by 2030. His research found this statistic originated from a Boston Consulting Group analysis that merely cited “experience with clients” as evidence – a claim Joshi describes as “flimsy.” Ironically, Google later admitted in its sustainability report that AI development was significantly increasing its corporate emissions.

Joshi’s new report, examining over 100 claims about AI’s climate benefits, discovered that only a quarter were backed by academic research, while more than a third cited no evidence whatsoever. This pattern of making unsubstantiated claims extends across the tech industry, with executives like OpenAI’s Sam Altman promising AI will “fix” climate change without substantial proof.

Generative AI vs. Specialized Machine Learning

The report highlights a critical distinction that tech companies often blur: the difference between energy-intensive generative AI (like ChatGPT) and more specialized, efficient machine learning systems. Nearly all claims examined conflated these different AI types, using examples of specialized AI’s benefits to justify the massive energy consumption of generative models.

David Rolnick, chair of Climate Change AI, notes that while specialized machine learning is already helping fight climate change in various sectors, tech companies are “relying on hypothetical AI that does not exist now” when making climate claims about generative models. This creates what researcher Sasha Luccioni calls a false narrative that “we need big AI models – and quasi-infinite amounts of energy.”

The Energy Impact of AI Development

The AI race is having tangible climate impacts. In the US, data center expansion has resulted in coal plants staying open and hundreds of gigawatts of new gas power planned – with nearly 100 gigawatts specifically for data centers. Meanwhile, tech companies remain opaque about their AI systems’ energy usage, making independent assessment difficult.

Luccioni and fellow researcher Yacine Jernite found that smaller, more efficient AI models often perform just as well as massive ones for specific applications. “The only companies that can compete in this bigger-is-better AI race are the ones with the deepest pockets,” Luccioni notes.

The Path Forward

Experts argue that greater transparency is essential. Joshi suggests tech companies should disclose specific energy usage data for their AI systems: “If they’re worried that people are overstating the climate impacts of generative AI, then there should be nothing stopping them from saying, ‘Our energy growth this year was six terawatt-hours, and two of them were for generative AI.'”

Without such transparency, it remains difficult to evaluate whether the climate costs of the AI boom will be outweighed by its potential benefits – or if tech companies are simply using vague climate promises to justify their energy-intensive expansion.