Since the launch of ChatGPT in November 2022, there has been a remarkable surge in the investment, development, and use of artificial intelligence (AI) applications worldwide. This rapid growth has not only had social and economic impacts but also environmental consequences that have received less attention.
The energy demand of AI applications is staggering. A single query to an AI-powered chatbot can consume up to ten times more energy than a traditional Google search. In general, a generative AI system may require 33 times more energy to complete a task compared to traditional software. This increased energy demand leads to higher carbon emissions and water usage, further straining electricity grids already affected by climate change.
Most AI applications run on servers in data centers. Even before the AI boom, data centers accounted for 1-1.5% of global electricity use and around 1% of energy-related CO? emissions. To put this into perspective, the aviation sector accounted for 2% of global energy-related CO? emissions in 2022, while the steel sector was responsible for 7-9%.
The rapid growth in AI use is changing these figures. Companies like Microsoft, Meta, and Google have disclosed their environmental impact. Microsoft’s annual emissions increased by around 40% between 2020 and 2023, reaching 17.1 million tonnes of CO?. Meta’s Scope 3 emissions increased by over 65% in just two years, reaching 8.4 million tonnes of CO? in 2022. Google’s emissions were almost 50% higher in 2023 compared to 2019.
Water consumption is another concern. Data centers require large amounts of water for cooling purposes. In the United States, data centers use about 7,100 liters of water for each megawatt-hour of energy consumed. Google’s US data centers alone consumed an estimated 12.7 billion liters of fresh water in 2021. This becomes particularly problematic in regions where water stress is increasing due to climate change.
The location of data centers can also pose risks. In times of scarcity, data centers may compete with local populations for resources, such as during extreme heat events. The increasing number of days above 50°C globally raises health risks for the population. Data centers’ power usage spikes with temperature, increasing the risk of blackouts and instability in electricity grids.
Tech companies are becoming more aware of these issues and taking action. Some have committed to replenishing more water than they consume by 2030 and are implementing initiatives to improve water security. However, there is still a lack of detailed sustainability data provided by data center operators, according to a survey of Australian sustainability professionals.
IT managers are also concerned about the increased energy consumption driven by AI applications but lack the necessary skills to address these sustainability impacts. Education and training for IT managers to understand and tackle the environmental consequences of AI is urgently needed.
