Whether in automotive engineering, retail, web navigation or road traffic - artificial intelligence (AI) is already being used in many areas of the economy and society. In the process, this key technology is becoming increasingly powerful and successfully opening up more and more areas of application. AI can also make an important contribution in the fight against climate change. But how exactly is this possible? How do AI and energy conservation fit together when the application and energy consumption of AI systems continues to rise?
In a new episode of the science podcast "Neuland," AI experts Professor Ralf Herbrich, managing director of the Hasso Plattner Institute (HPI) and head of the "AI and Sustainability" department at HPI, and Professor Robert Williamson, professor of "Fundamentals of Machine Learning" at Eberhard Karls University in Tübingen, talk about how AI can be used in the fight against climate change and how AI can become more energy efficient.
Accounting for about 4 percent of global energy consumption, data centers are among the major energy consumers. Training AI models is a relevant factor here, because the more powerful AI systems become, the more energy they require for their complex calculations. These are predictive calculations that are needed for decision-making in real-world applications. If we add to that the use of digital technologies such as laptops or smartphones, this accounts for nearly 8% of energy consumption. "And that number is not going to go down. Our daily lives are becoming more convenient, and our industries are becoming more efficient and productive through the use of such technologies," says Herbrich.
At the same time, AI systems have great potential to aid us in achieving climate goals. Artificial intelligence is already being used to make households more energy efficient through smart light switches and intelligent thermostats. AI has also been used to make progress in transportation in terms of safety and handling. The use of AI in solar energy storage is also conceivable. "You could name 10-20 other areas where AI technology can be used. Which one ends up being the most important one you’ll only know when you’ve tried it out. (...) There will be solutions, we just don't know what they look like yet," Williamson says.
Profound knowledge about the digital world, explained in a clear and understandable way – that’s what the science podcast "Neuland" with experts from the Hasso Plattner Institute (HPI) offers at: https://podcast.hpi.de, on iTunes and Spotify. Once a month on Neuland, they talk about current and socially relevant digital topics, their research work, and the opportunities and challenges of digital trends and developments.