Artificial Emotional Intelligence – A Game Changer for AI and Society?

Björn Schuller

It is with emotional intelligence that we recognise the emotions and feelings of those around us and ourselves. And it is emotion that pilots the way we behave and think – ultimately, that makes us attain our aims. In fact, emotion is broadly considered to be a survival factor for humanity. So where is today's Artificial Intelligence in terms of emotional intelligence? And how will Artificial Emotional Intelligence shape future AI, and in the longer run society? Here, I first give insight into the state of the art in machine recognition of human emotion from audio, video, and bio-signals. This includes recent trends in efficient dynamic active cooperative learning to exploit "big" social media data for deep end-to-end learning alongside the latest benchmarks in the field. The combination with multimodal emotion synthesis is shown by examples of current and past emotional agents. I will then discuss emotions in machine learning and decision making before a conclusion that will target the ethical, legal, and social implications, once machines broadly start to recognise, show, and potentially have emotions guide their learning, planning, and actions.

Björn Schuller is: