Potential and challenges of AI in sustainability
Sean Jones opened the discussion optimistically: “AI has the potential to solve problems and save the world.” Focus topic of this TUM Think Tank panel discussion was the use of AI in sustainability, the potential and the challenges. Lisa Klepper, IT Architect at Stadtwerke München, Prof. Stephan Günnemann, Executive Director of the MDSI, and Sean Jones, Chief Sustainability Officer at Microsoft discussed on ‘Beyond the Hype: How can we use AI to work towards sustainability goals and what can we do to co-design more ecological and social technologies?’.
One important aspect in the discussion was the selection of meaningful use cases and projects in which a sustainable benefit can be achieved. AI can successfully support in several topics. Stephan Günnemann gave several examples: “Regarding our environmental footprint, AI can help for instance to identify and design new materials with improved properties in energy generation. Other big chances lie in better drug design, health products or optimization of infrastructure.” Lisa Klepper added another aspect, AI can also improve the decision-making process compared to traditional, experience-based processes, but ethical and political aspects must also be taken into account.
Key to all applications: you always have to weigh up whether the use of AI justifies the high energy costs und carbon footprint it generates. As a closing remark, Sean Jones ended with an important message: “AI is not a threat and you [the audience] are the future: use AI responsibly for the benefit of the humankind.”
Sustainable energy conversion and storage
The climate change is real. Prof. David Egger emphasized ‘The Importance of New Materials for Sustainable Energy Conversion and Storage’, the topic of his talk, with impressive pictures of a vanishing glacier in Vorarlberg and the Danube flood.
Opening the TUM Venture Labs Talk Series ‘One Afternoon. Endless Learning Opportunities: A Sustainability Day Talk Series’, in his keynote lecture David Egger explained the need for new or optimized materials for solar cells. His group focuses on theoretical modelling to reach this goal. One main concern is the replacement of toxic elements which are present in certain solar cells. In general, within the last years, the cost of solar cells dropped so the new focus lies on increasing the efficiency, ideally with sustainable materials. Another important aspect is energy storage, Egger and his team work on solid state batteries and again there are many opportunities to improve the state-of-the-art. There is a lot of research ahead and David Egger ended the discussion with a personal note: “As a theorist one needs a high tolerance for frustration, but it is worth it.”
Meeting the experts
At the MDSI booth, two PhD students of the Group for Theory of Functional Energy Materials (TheoFEM) explained in further detail what Prof. David Egger presented in his talk. They shared first-hand experiences on how to apply computational techniques such as machine learning with neuronal networks to reveal and describe functional energy materials for solar cells.
The two founders of Pruna AI also engaged in lively discussions at the MDSI booth and presented their approach to reduce AI costs, times, memory and carbon emissions by a factor of 2 to 10.