idw – Informationsdienst Wissenschaft
Nachrichten, Termine, Experten
Nachrichten, Termine, Experten
The use of artificial intelligence (AI) is currently growing rapidly. This growth is accompanied by an increasing energy demand, rising greenhouse gas emissions and higher water and resource consumption. On behalf of Greenpeace Germany, Oeko-Institut has analysed the environmental impact of artificial intelligence and prepared a trend analysis for up to 2030. This report shows the ways in which AI can become more sustainable and formulates policy options for reducing its harmful environmental impact.
AI data centres drive consumption
As the use of AI increases, digital infrastructures – especially AI-specific data centres – are expanded. According to forecasts, the global electricity consumption of AI data centres will increase eleven-fold from the reference year of 2023 to 2030: from 50 billion kilowatt hours to around 550 billion kilowatt hours. Along with the other data centres, this means that around 1,400 billion kilowatt hours of electricity will be used for central data processing in 2030.
This is associated with an increase in greenhouse gas emissions from data centres from 212 million tonnes in 2023 to 355 million tonnes in 2030, despite the assumed expansion of the use of renewable energies for electricity production.
Other environmental burdens include the water required for cooling, which will almost quadruple to 664 billion litres, and the additional electronic waste generated by the expansion of data centres and AI capacity, which will amount to up to 5 million tonnes, over the same period. In addition, 920 kilotonnes of steel and around one hundred kilotonnes of critical raw materials will be required.
Renewable energies do not cover estimated demand
With the rising demand for energy, local power grids are increasingly reaching their limits. The result, says Jens Gröger, Research Coordinator for Sustainable Digital Infrastructures at Oeko-Institut, is that ‘in the years ahead, data centres will continue to rely on fossil fuels such as natural gas and coal – with correspondingly high environmental costs.’ The major technology companies are also investing in nuclear power plants and small modular reactors (SMRs).
Indirect effects of AI cannot be neglected
In addition to the so-called ‘direct’ environmental impacts – greenhouse gas emissions from data centres, water consumption for cooling systems and the use of resources in hardware production – the report also highlights the environmental significance of the indirect and systemic effects of AI. It can already be observed that AI is being used to accelerate environmentally harmful business practices. For example, AI tools are being used to tap into new fossil energy sources faster and more effectively, to increase the intensification of monocultures and to further increase private consumption. At the same time, negative environmental effects can also arise unintentionally due to errors in the data basis, incorrect training, or the operation of AI systems. Such indirect effects are environmentally significant but have rarely been recognised and discussed to date.
Policy recommendations: to enshrine environmental risks in law and counteract them
To counter the risks, the study recommends the following measures at a policy level:
- The introduction of binding transparency and accountability requirements for providers of data centres and AI services, including the collection and publication of key figures relating to data centres, the introduction of an efficiency label for data centres and key figures on their environmental footprint that are specific to AI services.
- Ensuring grid integration and adaptation to renewable energy generation volumes by covering the loads at suitable times with capacities from clean energy or with their own battery storage systems.
- An update of the legal framework to take into account the environmental impact of artificial intelligence. This includes, for example, an impact assessment that provides for a structured and specific environmental assessment of AI systems.
Jens Gröger
Research Coordinator for Sustainable Digital Infrastructures &
Senior Researcher in the Sustainable Products & Material Flows Division
Oeko-Institut e.V., Berlin office
Phone: ++49 30 405085-378
Email: j.groeger@oeko.de
Dr. Peter Gailhofer
Research Coordinator for Digital Ethics and Governance &
Senior Researcher in Environmental Law & Governance Division
Oeko-Institut e.V., Berlin office
Phone: ++49 30 405085-352
Email: p.gailhofer@oeko.de
https://www.oeko.de/en/publications/environmental-impacts-of-artificial-intellig...
Oeko-Institut’s study ‘Environmental impacts of artificial intelligence’
Merkmale dieser Pressemitteilung:
Journalisten
Energie, Informationstechnik, Umwelt / Ökologie
überregional
Forschungsergebnisse
Englisch
Sie können Suchbegriffe mit und, oder und / oder nicht verknüpfen, z. B. Philo nicht logie.
Verknüpfungen können Sie mit Klammern voneinander trennen, z. B. (Philo nicht logie) oder (Psycho und logie).
Zusammenhängende Worte werden als Wortgruppe gesucht, wenn Sie sie in Anführungsstriche setzen, z. B. „Bundesrepublik Deutschland“.
Die Erweiterte Suche können Sie auch nutzen, ohne Suchbegriffe einzugeben. Sie orientiert sich dann an den Kriterien, die Sie ausgewählt haben (z. B. nach dem Land oder dem Sachgebiet).
Haben Sie in einer Kategorie kein Kriterium ausgewählt, wird die gesamte Kategorie durchsucht (z.B. alle Sachgebiete oder alle Länder).
