Artificial intelligence’s power-hungry nature has earned it a reputation as a climate pariah. Training a single model can use more electricity than 100 US homes in an entire year. As the world’s demand for AI grows, so will data centers’ energy consumption, which is projected to more than double by 2030, according to the International Energy Agency. That will have a material impact on the energy transition: A new report from BloombergNEF estimates that over the next decade, two-thirds of the additional power needed for data centers will come from fossil fuels. While concerns about AI are well-earned, the technology also has a role to play in fighting climate change. Finding ways to cut carbon emissions often requires analyzing vast datasets and identifying new patterns — something computers do better and faster than humans. Algorithms can also help communities better prepare for climate-induced risks, for instance, by combing through troves of images captured by cameras in fire-prone areas to spot wildfire outbreaks before they become obvious to the human eye. Scaling up existing AI solutions could cut more emissions than all the data centers generate, the IEA estimates. Still, AI “is not a silver bullet” for the global climate crisis, the group says.

Full commentary : A look at ways AI is being used to cut carbon emissions: calculating fuel-efficient driving patterns, monitoring grid infrastructure, and more.