Electricity guzzler AI: How ChatGPT pollutes our climate!

Electricity guzzler AI: How ChatGPT pollutes our climate!

The use of artificial intelligence (AI) poses ever greater ecological challenges. According to OpenAI, a request to ChatGPT, one of the best-known AI platforms, uses as much electricity as one second of operating an oven. OpenAI boss Sam Altman sees this as both a problem and an opportunity for the future of the technology. He expresses optimism about the role of AI in a more prosperous future, despite concerns about possible job losses.

Some figures illustrate the high resource consumption of AI systems. The water consumption for a single request is only a fifth of a teaspoon. Nevertheless, this consumption adds up enormously when you consider the large number of daily requests. Loud daily news ChatGPT-3 training requires an estimated 5.4 million liters of water, of which 700,000 liters are used to cool the data centers.

The growing energy demand

Technology companies Microsoft, Google and Amazon already have strategic plans to meet their increasing energy needs. Among other things, they rely on nuclear energy in order not to increase their carbon dioxide emissions. The energy requirements of data centers supporting AI applications is a big issue. A worrying trend shows that electricity consumption in data centers in Germany increased by 70 percent from 2010 to 2021. This is slower than the increase in demand for these facilities as servers become more efficient.

Another consideration is water consumption, which is essential for cooling the servers. By 2030, water consumption for server cooling is expected to reach 664 billion liters, almost four times as much as in 2023. The key here is the operation of air and water cooling systems, which require significant amounts of water.

Environmental policy and sustainability

The increase in the use of AI is leading to a worrying increase in greenhouse gas emissions. Forecasts show that emissions will rise from 212 million tonnes to 355 million tonnes by 2030. In order to combat this negative development, companies and political decision-makers are called upon to take action. On behalf of Greenpeace Germany, the Öko-Institut recommends binding transparency requirements and an efficiency label for data centers.

A future-oriented approach could be the integration of data centers into renewable energy and heating networks. The aim is to ensure that AI developments do not hinder climate protection, but rather actively contribute to the energy transition. However, the debate about the sustainability of AI is only just beginning, warns Ingenieur.de.
A conscious political and social framework must be created in order to make optimal use of the opportunities offered by AI for climate protection and to minimize the risks.

Given the challenges posed by the increasing energy and water consumption of AI data centers, there is an urgent need for further studies to measure environmental impacts more precisely and to develop measures to reduce them. The discussion about AI and its environmental consequences will continue to gain in importance.