Generative AI and climate change are on a collision course

In 2025, AI and climate change, two of the biggest social disruptors we face, will collide.

The summer of 2024 beat the record for the hottest day on Earth since data collection began, sparking media coverage and public debate. This is also the year that the two Microsoft and Googletwo of the big tech companies investing heavily in AI research and development have missed their climate targets. While this has also made headlines and spurred outrage, the environmental impacts of AI are still far from common knowledge.

In reality, the current “bigger is better” paradigm of AI—epitomized by tech companies’ pursuit of ever-bigger and more powerful language models that are presented as the solution to every problem—comes with very significant costs for the environment. These range from generating colossal amounts of energy to power the data centers that run tools like ChatGPT and Midjourney to the millions of gallons of fresh water that are pumped through these data centers to ensure they don’t overheat and tons of rare earth metals. need to build the hardware they contain.

Data centers already use it 2 percent of the electricity in the world. In countries such as Ireland, this figure reaches up to a fifth of the electricity generated, which prompted the Irish government to declare a effective moratorium on 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 credits of renewable energy, which technically compensate for the emissions incurred by the generation of this electricity, but don’t change the way in which it is generated.

Places like Data Center Alley‘ in Virginia are mostly powered by non-renewable energy sources such as natural gasand energy providers are delaying the retirement of coal plants to keep up with the increased demands of technology such as AI. Data centers are slurping up huge amounts of fresh water from scarce aquifers, pitting local communities against data center providers in locations ranging from Arizona to Spain. In Taiwanthe government has chosen to allocate precious water resources to chip manufacturing facilities to stay ahead of rising demands instead of letting local farmers use it to irrigate their crops amid the worst drought the country has seen in more than a century.

My latest research shows that moving from older standard AI models – trained to do a single task like answering questions – to the new generative models can use up to 30 times more energy just to answer the same set of questions. Tech companies that are adding more and more generative AI models to everything from search engines to word processing software have yet to disclose the carbon cost of these changes – we don’t even know how much energy is used during a conversation with ChatGPT or when generating. an image with Gemini from Google.

Much of the talk from Big Tech around the environmental impacts of AI has followed two trajectories: Either it’s not really a problem (according to Bill Gates), or an energy breakthrough will come and magically solve things (according to Sam Altman). What we really need is more transparency around the environmental impacts of AI, through voluntary initiatives like the AI Energy Star project I led, which helped users compare the energy efficiency of AI models to make informed decisions. He predicts that by 2025, voluntary initiatives like these will begin to be enforced 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. The environmental footprint of AI and take the necessary actions to reduce.