Published: 23/04/2025
The power of artificial intelligence (AI) to process huge amounts of data to help improve efficiency, make decisions and solve complex problems is transforming industries. The applications are increasingly being interwoven into our personal and work lives, ranging from smartphone GPS navigators to providing personalised services such as ChatGPT. According to one survey, nearly three out of four businesses already use AI for at least one business function. As such, the AI market is projected to grow from US$150.2 billion in 2023 to US$1.34 trillion by 2030.1
However, as AI’s popularity continues to rise, attention is being drawn to the environmental cost it poses. Training and running an AI system requires significant computing power and electricity. And the more sophisticated the application is, the more energy-intensive it becomes. A single query via ChatGPT reportedly takes ten times the energy of a Google search.2 This saw Google drop its long-standing carbon-neutral promise in 2024 following a spike in emissions, which rose nearly 50% since 2019. To help bring down emissions, tech giants like Microsoft, Google and Amazon are turning to alternative power sources like nuclear to power their data centres (which house and train the massive artificial intelligence models behind the generative AI applications).
AI also consumes a significant amount of water, both during its creation and for its daily maintenance. One study estimates that a two-week training for the GPT-3 AI program in Microsoft's U.S data centres consumes around 700,000 litres of clean freshwater, equivalent to producing 320 Tesla electric vehicles.3
Despite this, AI has the potential to be a valuable tool in the fight against climate change - from optimising power grids and enabling smart buildings to enhancing transportation systems and facilitating precision agriculture. AI’s predictive capabilities also have the potential to help tackle environmental emergencies from detecting patterns in data and predicting future outcomes to the development of mitigation policies.
During the recent COP29 climate summit, industry leaders, including AI hardware and software developer NVIDIA and German utility group EON, discussed “Green AI” which incorporates solutions to optimising energy. This includes developing lighter and more energy-efficient hardware, adopting techniques that can reduce AI training times, and processing data locally (as opposed to transmission to the cloud). The discussion also touched on the use of direct-to-chip liquid cooling to help data centres reduce water demands. By combining innovation with sustainability, Green AI is better placed to meet the growing demand for computational power while reducing its impact on the environment.
- Global Artificial Intelligence (AI) Market Research 2023-2030: Insights, Case Studies, Value Chain, Ethics & Implications, Best Practices, and More
- Tech and AI are changing the climate equation – for the worse | Lowy Institute
- AI programs consume large volumes of scarce water | UCR News | UC Riverside