Generative AI and climate change are on a collision course

In 2025, artificial intelligence and climate change, two of the biggest societal disruptions we face, will collide.

Summer 2024 He broke the record It was the hottest day on Earth since data collection began, sparking widespread media coverage and public debate. This also happens to be the year they both come together Microsoft and GoogleTwo leading major technology companies that invest heavily in artificial intelligence research and development have failed to meet their climate targets. While this has also made headlines and sparked outrage, the environmental impacts of AI are still far from being common knowledge.

In fact, the current “bigger is better” model of AI – exemplified by technology companies’ pursuit of larger, more powerful language models that are presented as a solution to every problem – comes at very high environmental costs. These range from generating massive amounts of power to power the data centers that power tools like ChatGPT and Midjourney to millions of gallons of fresh water pumped through these data centers to make sure they don’t overheat and tons of rare earth minerals. needed to build the devices that contain them.

Data centers are already being used 2% of global electricity. In countries such as Ireland, this figure amounts to a fifth of the electricity generated, prompting the Irish government to announce Effective stop in new data centers until 2028. While much of the energy used to power data centers is officially “carbon neutral,” this relies on mechanisms such as renewable energy credits, which technically offset the emissions generated by generating this electricity, but they do not change the way with which it was created.

Places like Data Center AlleyIn Virginia, it is powered mostly by non-renewable energy sources such as… Natural gasEnergy providers are delaying the retirement of coal-fired power plants to keep up with the times Increased demands Technologies such as artificial intelligence. Data centers consume huge amounts of fresh water from scarce aquifers, pitting local communities against data center providers in places ranging from Arizona to Spain. in TaiwanThe government has chosen to allocate precious water resources to build manufacturing facilities to stay ahead of growing demands rather than allow local farmers to use it to water their crops amid the worst drought the country has seen in more than a century.

My recent research shows that shifting from old standard AI models – trained to do a single task such as answering questions – to new generative models could use up to 30 times Just more energy to answer the exact same set of questions. Tech companies that are increasingly adding generative AI models to everything from search engines to word processing programs also aren’t disclosing the carbon cost of these changes — we still don’t know how much energy is used during a conversation with ChatGPT or when generating power. Image with Google Gemini.

Much of Big Tech’s rhetoric about the environmental impacts of AI has followed two paths: Either it’s not really a problem (according to… Bill Gates), or an energy breakthrough will come along and magically fix things (according to… Sam Altman). What we really need is more transparency about the environmental impacts of AI, through voluntary initiatives like this one Ai Energy Star The project I am leading, which will help users compare energy efficiency in AI models to make informed decisions. I expect that in 2025, such voluntary initiatives will begin to be implemented through legislation, from national governments to intergovernmental organizations such as the United Nations. In 2025, with more research, public awareness, and regulation, we will finally begin to understand this The environmental footprint of artificial intelligence And take the necessary measures to reduce it.