Report questions Big Tech claims that AI can help fight climate change

Technology companies are increasingly conflating traditional artificial intelligence (AI) with generative AI while promoting the sector’s climate benefits, a new analysis has claimed, arguing that many such assertions lack verifiable evidence of emissions reduction.

While traditional AI uses data-driven models to analyse patterns, make predictions or automate specific tasks — such as fraud detection or weather forecasting. Generative AI, by contrast, creates new content like text, images or video using large language models, typically requiring far more computing power, data processing and energy-intensive infrastructure.

The research, commissioned by non-profits Beyond Fossil Fuels and Climate Action Against Disinformation, examined 154 public statements and found that most claims about AI helping avert climate breakdown referred to machine learning or predictive tools rather than energy-intensive chatbots and image-generation systems driving rapid expansion of datacentres.

According to the report, it did not identify a single instance where widely used generative AI tools — such as Google’s Gemini or Microsoft’s Copilot — resulted in a “material, verifiable, and substantial” reduction in greenhouse gas emissions.

Energy analyst Ketan Joshi, who authored the report, alleged that parts of the industry’s messaging amounted to “greenwashing”. He said companies were highlighting smaller climate initiatives while expanding infrastructure that significantly increases energy demand.

“These technologies only avoid a minuscule fraction of emissions relative to the massive emissions of their core business,” Joshi said, comparing the approach to fossil fuel firms promoting limited renewable investments.

The analysis reviewed claims cited in an International Energy Agency (IEA) report and corporate disclosures from major technology firms. It said only about 26 per cent of climate-related claims referenced published academic research, while more than a third did not cite supporting evidence.

The report was released during the AI Impact Summit held in Delhi this week and argued that the industry had “muddled” different forms of AI, presenting climate solutions alongside rising carbon emissions from datacentre growth.

Sasha Luccioni, AI and climate lead at Hugging Face, who was not involved in the research, said debates around AI often grouped very different technologies together. She noted that generative AI and large language models tend to be more energy-intensive, while older predictive models are more commonly linked to climate mitigation use cases.

Industry estimates suggest datacentres currently consume about 1 per cent of global electricity, but their share of electricity use in the United States could rise sharply over the next decade due to demand from AI workloads. The IEA has projected that datacentres could account for at least one-fifth of electricity demand growth in advanced economies by the end of the decade.

While companies argue that AI can optimise energy systems and improve efficiency, critics say the climate impact of large-scale datacentre expansion requires closer scrutiny. A Google spokesperson said its emissions reduction estimates were based on scientific methodologies and transparent reporting principles. Microsoft declined to comment, while the IEA did not respond to requests for comment.

Joshi said public discourse on AI’s environmental role needed to be “brought back to reality”, warning that overstating benefits could distract from the environmental costs associated with rapid technological expansion.