Artificial intelligence: water consumption is exploding - are we becoming poor?

Artificial intelligence: water consumption is exploding - are we becoming poor?

The rapid increase in the use of artificial intelligence (AI) presents not only remarkable progress but also significant ecological challenges. Studies show that data centers supporting AI applications require not only large amounts of electricity but also water for cooling. This urgent concern has been highlighted in various analyses, for example in a report by crown, which addresses the complex relationships between AI, energy consumption and water requirements.

OpenAI boss Sam Altman expressed optimism about the future of AI in a blog post, but acknowledged the severe social disruption that could accompany the rise of these technologies. Altman predicts that artificial intelligence will make the world richer, which could fund new policy ideas such as a possible basic income fueled by productivity gains. But despite these positive outlooks, the environmental impact of AI applications is significant.

Environmental footprint of AI and data centers

The water and energy consumption data is alarming. Loud daily news Water consumption from a single conversation with a chatbot can be up to 500 milliliters. The development of AI models like ChatGPT-3 required an estimated 5.4 million liters of water, of which 700,000 liters were used to cool the data centers alone. Water demand could increase dramatically in the future: it is predicted that a total of 664 billion liters of water will be needed to cool servers by 2030.

The energy requirements of data centers are also increasing rapidly. Electricity use will increase from 50 billion kilowatt hours in 2023 to around 550 billion kWh in 2030. This development will lead to an increase in greenhouse gas emissions, which are forecast to rise from 212 million tonnes in 2023 to 355 million tonnes in 2030. Such developments could seriously jeopardize climate goals, as the analysis by the Öko-Institut on behalf of Greenpeace Germany shows.

Sustainability and political measures

Dependence on fossil fuels remains problematic as many data centers continue to rely on conventional energy sources. According to reports from Ingenieur.de By 2030, the total energy requirement for data processing will rise to around 1,400 billion kWh. This urgently requires political action patterns to systematically assess and regulate the environmental impacts of AI use.

Recommended measures include mandatory transparency requirements for data center providers, the development of an efficiency label for these facilities, and the integration of renewable energy into the cooling and power supply of these centers. Politicians have a responsibility to create framework conditions that ensure that the innovation potential of AI does not come at the expense of the environment and climate protection.