Artificial intelligence (AI) has drawn criticism for its significant energy and water consumption. Yet scientists are exploring ways it can help people and businesses use energy more efficiently and reduce pollution.
According to the International Energy Agency, data centres powering AI accounted for around 1.5% of global electricity consumption last year, and this figure could more than double by 2030. Higher energy use may increase reliance on fossil fuels like coal and gas, which emit greenhouse gases that contribute to global warming, rising sea levels, and extreme weather.
However, AI can also analyse energy use and pollution to make buildings more efficient, optimise device charging, reduce emissions from oil and gas operations, locate geothermal resources, and manage traffic to cut vehicle emissions. Experts say these applications could help offset the energy consumed by AI.
“I am optimistic that even as AI use grows, advances in processing efficiency will prevent energy consumption from rising as sharply as predicted,” said Alexis Abramson, dean of Columbia University’s Climate School.
Building efficiency: Maintenance and cooling
AI can optimise building energy use by automatically adjusting lighting, ventilation, heating, and cooling based on weather, occupancy, and electricity data, said Bob French, chief evangelist at 75F, a building automation company. US homes and buildings contribute roughly one-third of the country’s greenhouse gas emissions.
Automated thermostats and ventilation systems schedule heating and cooling efficiently, reducing energy waste and preventing costly repairs by predicting equipment failures. Combined, these measures can lower building energy consumption by 10 to 30%.
Efficient EV and device charging
AI can schedule electric vehicle (EV) charging and device usage for periods of lower grid demand or higher renewable energy availability, reducing fossil fuel reliance and saving money. In California, pilot programmes have demonstrated cost savings by shifting charging to optimal times. AI can also help homeowners with solar panels store excess energy efficiently.
Reducing methane emissions in oil and gas
Boston-based Geminus AI uses deep learning to minimise methane flaring and venting, a key contributor to global warming. By simulating pipeline and compressor operations, AI can identify changes to reduce emissions in seconds — a process that traditionally takes engineers 36 hours. “Scaling this across the industry offers a huge opportunity to cut greenhouse gas emissions,” said CEO Greg Fallon.
Identifying geothermal resources
Geothermal energy startup Zanskar uses AI to model Earth’s subsurface, locating overlooked geothermal hot spots for clean electricity generation. AI simulations help choose optimal drilling sites, unlocking resources that would be difficult to identify manually. Last year, Zanskar revived an underperforming New Mexico plant and made a new discovery in Nevada, demonstrating AI’s potential in renewable energy expansion.
Cutting traffic emissions
Google’s Project Green Light uses AI and Maps data to optimise traffic light timing, reducing stop-and-go traffic and vehicle emissions. Passenger cars and small trucks contribute about 16% of US greenhouse gas emissions. The programme, now in 20 cities globally, can cut stop-and-go traffic by 30%, lowering emissions by 10% while improving air quality.
“We are only beginning to explore AI’s potential,” said Juliet Rothenberg, Google’s product director for Earth and resilience AI.
AI’s high energy use is a challenge, but its applications in efficiency, renewable energy, and emission reduction show it can play a key role in protecting the planet.