AI’s climate problem is worse than we thought

Key Points:

• Recent discussions on AI and climate change have focused mainly on the electricity consumption of data centers powering AI, which has significantly increased fossil fuel use and emissions from tech companies.
• New research reveals that AI's role in enabling fossil fuel companies to extract and produce more oil and gas results in emissions 3.3 to 13.3 times greater than those from powering AI data centers, potentially equating to the annual emissions of large countries like Mexico or Russia.
• The study shows that AI's benefits to renewable energy are insufficient to offset the emissions enabled by its use in fossil fuel production unless renewables gain a productivity boost four to five times greater than fossil fuels.
• Experts Holly and Will Alpine emphasize the need to expand the climate conversation beyond operational emissions to include "enabled emissions," calling for governance and corporate accountability on how AI is used in fossil fuel expansion.
• Current policy efforts addressing AI's climate impact remain limited and focused on data centers, with voluntary standards and preliminary proposals; the Alpines advocate for mandatory disclosures and restrictions on AI tools designed to increase fossil fuel production.