Study Calls AI Climate Promises “Greenwashing”
A new report analyzing claims by major technology companies about the role of artificial intelligence (AI) in combating climate change has found many of those claims lack solid evidence and may mislead the public. Researchers studied 154 claims that described AI’s environmental advantages and determined that most of them contained ambiguous information which had not been confirmed and which described technological advances that existed before the current AI development phase.
The research, conducted by an energy analyst and commissioned by environmental groups such as Beyond Fossil Fuels and Climate Action Against Disinformation, examined statements by big tech firms and industry reports. The study showed that none of the generative AI tools which power popular chatbots and image generators could produce substantial reductions in greenhouse gas emissions which could be verified through direct measurement. The claims used older machine learning methods which required less energy than current AI development uses.
Evidence Gaps and Misleading Messaging
The researchers discovered that only 26 percent of all examined claims included references to peer-reviewed academic studies. The study found that 36 percent of the claims made no attempts to present any proof. The world heard multiple times the claim that AI technology would decrease worldwide emissions by 5 to 10 percent before the year 2030. The estimate originated from corporate documents which had not been verified through independent research studies.
Analysts report that companies within the industry combine various AI technologies which include both classical machine learning and current generative systems to make it seem all AI systems help decrease environmental problems. The process which lets organizations develop large language models and other similar systems results in datacenter operations consuming more energy which leads to serious environmental problems. Current datacenter operations account for approximately 1 percent of worldwide electricity consumption, but experts project their usage will increase significantly during upcoming years.
The Role of Datacenters and Energy Consumption
The use of generative AI tools needs extensive processing power for their training and operational functions. The processing power used by these systems originates from data centers that must find the right balance between delivering optimal performance and consuming power. The electricity consumption for basic AI queries remains low. However, video generation requires a substantial increase in power usage which results in greater carbon emissions. Tech companies present their efficiency improvements and climate technology initiatives as evidence of AI’s climate change mitigation benefits. However, the actual impact of these initiatives remains unknown or results in negative outcomes.
Calls for Greater Transparency
Current industrial messaging needs better transparency requirements together with independent assessment of AI systems’ environmental effects according to critics of existing industry practices. The researchers demonstrate that AI systems building a larger infrastructure through their development will generate more emissions instead of producing climate benefits. The experts explain that using different AI systems in marketing materials results in misrepresentation of modern generative systems’ actual energy requirements which creates difficulties for both policymakers and public audiences when they attempt to evaluate climate change solutions.
