Tech companies are conflating traditional artificial intelligence with generative AI when claiming the energy-hungry technology could help avert climate breakdown, according to a report.
Most claims that AI can help avert climate breakdown refer to machine learning and not the energy-hungry chatbots and image generation tools driving the sector’s explosive growth of gas-guzzling datacentres, the analysis of 154 statements found.
The research, commissioned by nonprofits including Beyond Fossil Fuels and Climate Action Against Disinformation, did not find a single example where popular tools such as Google’s Gemini or Microsoft’s Copilot were leading to a “material, verifiable, and substantial” reduction in planet-heating emissions.
Ketan Joshi, an energy analyst and author of the report, said the industry’s tactics were “diversionary” and relied on tried and tested methods that amount to “greenwashing”.
He likened it to fossil fuel companies advertising their modest investments in solar panels and overstating the potential of carbon capture.
“These technologies only avoid a minuscule fraction of emissions relative to the massive emissions of their core business,” said Joshi. “Big tech took that approach and upgraded and expanded it.”
Most of the claims that were scrutinised came from an International Energy Agency (IEA) report, which was reviewed by leading tech companies, and corporate reports from Google and Microsoft.
The IEA report – which devoted two chapters to the potential climate benefits of traditional AI – had a roughly even split between claims that rested on academic publications, corporate websites and those that had no evidence, according to the analysis. For Google and Microsoft, most claims lacked evidence.
The analysis, released during the AI Impact Summit in Delhi this week, argues the tech industry has misleadingly presented climate solutions and carbon pollution as a package deal by “muddling” types of AI.
Sasha Luccioni, AI and climate lead at Hugging Face, an open-source AI platform and community, who was not involved in the report, said it added nuance to a debate that often lumped very different applications together.
“When we talk about AI that’s relatively bad for the planet, it’s mostly generative AI and large language models,” said Luccioni, who has pushed the industry to be more transparent about its carbon footprint.
“When we talk about AI that’s ‘good’ for the planet, it’s often predictive models, extractive models, or old-school AI models.”
Green claims even for traditional AI tended to rely on weak forms of evidence that had not been independently verified, the analysis found. Only 26% of the green claims that were studied cited published academic research, while 36% did not cite evidence at all.
One of the earliest examples identified in the report was a widespread claim that AI could help mitigate 5-10% of global greenhouse gas emissions by 2030.
The figure, which Google repeated as recently as April last year, came from a report it commissioned from BCG, a consulting firm, which cited a blogpost it wrote in 2021 that attributed the figure to its “experience with clients”.
Datacentres consume just 1% of the world’s electricity but their share of US electricity is projected to more than double to 8.6% by 2035, according to BloombergNEF. The IEA predicts they will account for at least 20% of the rich world’s growth in electricity demand to the end of the decade.
While the energy consumption of a simple text query to a large language model such as ChatGPT may be as little as running a lightbulb for a minute, partial industry disclosures suggest, it rises considerably for complex functions such as video generation and deep research, and has troubled some energy researchers with the speed and scale of its growth.
A spokesperson for Google said: “Our estimated emissions reductions are based on a robust substantiation process grounded in the best available science, and we have transparently shared the principles and methodology that guide it.”
Microsoft declined to comment, while the IEA did not respond to requests for comment.
Joshi said the discourse around AI’s climate benefits needed to be “brought back to reality”.
“The false coupling of a big problem and a small solution serves as a distraction from the very preventable harms being done through unrestricted datacentre expansion,” he said.
