2025 Essays

Using energy-hungry AI to detect climate tipping points is a paradox

Sathuluri, David; Tedesco, Marco

The article contends that the increasing use of highly energy-intensive AI systems to detect climate tipping points constitutes a paradox that risks undermining decarbonisation efforts. It argues that the rapid expansion of computationally demanding models and data centres significantly elevates electricity demand and associated emissions at a time when strict adherence to limited carbon budgets is imperative. In the absence of robust governance frameworks that constrain AI infrastructure within science-based climate targets, such deployment patterns may lock in long-lived fossil fuel assets and jeopardise the achievement of climate goals. The author therefore advocates for regulatory mechanisms, efficiency standards, and renewable energy requirements that explicitly align AI development with the objectives of the energy transition and climate justice.