International Symposium - Humans at the Centre of HRI, 2022: Laurence Devillers - Video

https://europe.naverlabs.com/research/1st-hri-international-symposium/

Affective robotics & nudge: ethical issues
Emerging interactive and adaptive systems using emotions such as affective robots modify how we will socialise with machines. These areas inspire critical questions centering on the ethics, the goals and the deployment of innovative products that can change our lives and society. The concept of nudge is already used in social robotics even
if the nature of the mechanisms that characterize it is not always consistent. Nudge in social robotics is being used positively in several domains: health, education, etc. However, AI-enhanced Nudges (i.e. all kind of digital nudges using AI mechanisms as statistical inferences from users’ behaviour) are rising new ethical concerns. A set of goals defined by human agents may be reached using decision-making mechanisms, recommendations or other interaction influences. In addition, some nudge mechanisms are built unintentionally by the system to achieve its ends. Nudging mechanisms can occur at a very fine level of granularity e.g. adaptation of the tone of voice of the robot based on online learning variants of machine learning. These processes may generate a number of ethical risks for individuals, groups or society.
Laurence Devillers is a full Professor of computer science applied to humanities and social sciences at Sorbonne University, head of the research team (5-10p.) on “Affective and social dimensions of spoken interactions with (ro)bots and ethical issues” at LISN-CNRS (Paris-Saclay). She also leads the AI chair: “HUman-MAchine AffectiveInteraction & Ethics” of the DATAIA Saclay Institute.