Abstract: Artificial Intelligence (AI) has become a pervasive technology in modern societies. Naturally, this has resulted in questions being raised regarding the ethical use of AI. However, a relatively under-studied aspect of modern AI is the relationship between AI and the environment. Used correctly AI has the potential to help our societies become move environmentally sustainable. At the same time the lifecycle of AI systems, and in particular the development and deployment of large Deep Learning models, has a direct environmental cost. In this talk, I will highlight the environmental cost of modern AI and describe some of the ongoing research that is attempting to make AI more environmentally sustainable.
Bio: Prof. John Kelleher is the academic leader of the Information, Communication, and Entertainment (ICE) research institute at TU Dublin. He is the TU Dublin lead for the Science Foundation Ireland funded ADAPT research centre (https://www.adaptcentre.ie) and PhD Centre for Research Training in Digitally Enhanced Reality (https://d-real.ie), and the H2020 project PRECISE4Q which focuses on using Artificial Intelligence to develop clinical decision support systems for the treatment of stroke (https://precise4q.eu). He has published three books on Machine Learning: Deep Learning (https://mitpress.mit.edu/books/deep-learning-1), Data Science (https://mitpress.mit.edu/books/data-science), and Fundamentals of Machine Learning for Predictive Data Analytics (https://mitpress.mit.edu/books/fundamentals-machine-learning-predictive-data-analytics-second-edition).