This talk will discuss the challenges of engineering sustainable AI using face and emotion recognition as case studies. Existing technologies are often based on simplistic emotion models and can lead to false or even sexist or racist conclusions. As a society, we need to remember that AI is not autonomous and “objective”. Machines can learn, but they learn what we teach them, and as humans, it is our duty to teach our algorithms responsibly. As consumers, we need to make informed decisions about AI technologies, especially those claiming to give us new insights into complex problems like emotion recognition.
About the Speakers
Dr Magdalena Rychlowska is a Lecturer in the School of Psychology at Queen’s University Belfast and an Honorary Fellow in the Department of Psychology at the University of Wisconsin-Madison. She earned her PhD in 2014 from the University of Clermont-Ferrand in France and joined Queen’s University in 2017. She teaches research methods and cross-cultural psychology. Her research focuses on emotion and social signals, in particular smiles and laughter which are among the most common yet most understudied human expressions. A lot of her time is spent trying to figure out what smiles there are, why some laughs and smiles are more positive than others, and what makes people laugh. In short, she has the best job in the world. When she does not work, she reads, cooks, walks, or talks to her family and friends.
Dr Gary McKeown is a Senior Lecturer in the School of Psychology, Queen's University Belfast. His area of expertise is Affective Computing. That is getting computers to simulate and be responsive to emotional and social signals that humans make when they communicate with one another. Affective computing uses sensor and camera technology, awareness of context and machine learning algorithms to recognise and interpret emotional and empathic behaviour. It can also use knowledge of human emotions to synthesise emotional and empathic behaviour in Avatars. Gary is an editor for the field’s flagship journal Transactions on Affective Computing and a member of the Executive Committee of the Association for the Advancement of Affective Computing. His research in psychology provides theoretical input and practical databases that inform the science of social interaction and emotions and produces data for affective computing algorithms.