Since the discharge of ChatGPT in November 2022, the world has seen an incredible increase in investment, development and use of artificial intelligence (AI) applications. According to 1 estimate, the computing power used for AI is doubling roughly every 100 days.
The social and economic impact of this boom has provoked reactions around the globe. European regulators recently pushed Meta is pausing plans to coach AI models using users’ Facebook and Instagram data. The Bank for International Settlements, which coordinates the world’s central banks, has warned The use of artificial intelligence could change the way in which inflation works.
The environmental impact has received less attention to this point. A single request to an AI-powered chatbot can have as much as ten times as much energy than old-fashioned Google search.
Generally speaking, a generative AI system could 33 times more energy to finish a task than it will with traditional software. This enormous energy demand results in a rise in carbon emissions and water consumption and might put further strain on power grids already strained by climate change.
energy
Most AI applications run on servers in data centers. In 2023, before the actual AI boom, the International Energy Agency estimated that data centers would already 1–1.5% of worldwide electricity consumption and around 1% of worldwide energy-related CO₂ emissions.
For comparison: In 2022, the aviation sector accounted for 2% of worldwide energy-related CO₂ emissions while the steel sector answerable for 7–9%.
How does the rapid growth of AI usage change these numbers? Recent environmental reports from Microsoft, Meta and Google provide insight.
Microsoft is investing heavily in AI and holds a big stake in ChatGPT manufacturer OpenAI in addition to its own Copilot applications for WindowsBetween 2020 and 2023, Microsoft will disclosed annual emissions increased by around 40% from 12.2 million tonnes of CO₂ comparable to 17.1 million tonnes.
These figures include not only direct emissions, but additionally indirect emissions, comparable to those generated when generating electricity to run data centers and people generated by utilizing the corporate's products. (These three emission categories are known as Scope 1, Scope 2 and Scope 3 emissions.)
Meta also drops enormous resources in AIIn 2023, the corporate announced that its Scope 3 emissions increased by over 65% in only two years from 5 million tonnes of CO₂ in 2020 to eight.4 million tonnes in 2022.
Google’s emissions were almost 50% higher in 2023 than in 2019. The tech giant's 2024 environmental report notes that planned emissions reductions shall be difficult “as a result of increasing energy demands brought on by the upper intensity of AI computing.”
Water
Data centers generate a whole lot of heat and use large amounts of water to chill their servers. According to a Study 2021Data centers within the USA devour about 7,100 liters of water per megawatt hour of energy.
Google’s US data centers alone consumed estimated 12.7 billion liters of fresh water in 2021.
In regions where climate change is exacerbating water shortages, water use by data centers is becoming a specific problem. The recent drought in California, where many technology corporations are based, has led to corporations like Google, Amazon And Meta to launch “water positive” initiatives.
These major technology corporations have announced their goal to replenish more water than they use by 2030. Their plans include projects comparable to designing ecologically resilient watershed landscapes and improving municipal water conservation to enhance water security.
Climate risk
When data centers are situated in or near cities, they might compete with the population for resources during times of scarcity, comparable to extreme heat events.
Worldwide, the entire variety of days with temperatures above 50°C elevated in every decade since 1980. July 2023 was the hottest month ever.
Extreme heat has health consequences for the local population. Lancet 2022 study found that even a temperature increase of 1°C was positively related to increased mortality and morbidity.
On days of maximum heat, air-con can save lives. Data centers also must be kept cool, so their power consumption increases as temperatures rise, increasing the danger of blackouts and instability in power grids.
What's next?
So what now? As we’ve got seen, technology corporations are increasingly aware of the issue. How is that this being put into motion?
If we Australian sustainability experts interviewed In July 2023, we found that only 6% of respondents believed that data center operators provide detailed sustainability data.
At the start of the yr we IT managers surveyed in Australia and New Zealand to ask them what they consider how AI applications are boosting energy consumption. We found that 72% are already adopting or testing AI technologies.
More than two-thirds (68%) said they were concerned in regards to the increasing energy consumption for AI requirements, but there may be also considerable uncertainty in regards to the magnitude of the rise.
Many IT managers also lack the mandatory skills to adequately consider these sustainability impacts, no matter corporate sustainability commitments. There is an urgent need to coach and train IT managers to know and consider the sustainability impacts of AI.
