Paper short abstract:

Starting from our experience with culture-aware robotics, the article introduces the concept of a "diversity-aware robot", i.e., an embodied AI that takes into account the unique characteristics of the person it is interacting with to be more easily accepted and provide them with better services.

Paper long abstract:

In the last five years, the project CARESSES introduced the new concept of transcultural robotic nursing. According to this new paradigm, Socially Assistive Robots shall not only be capable of performing useful services (e.g., to provide access to technology, foster social participation, engage people in activities, reduce loneliness, etc.), but also be "culture-aware" and take into account the cultural context and the background of the person they are interacting with. To this end, the robot's Artificial Intelligence (AI) must be designed in order to avoid stereotypical representation of people: the AI must be informed by our a priori knowledge about cultures but also be able to acquire information about individual characteristics during the interaction.

The article discusses how culture-aware robotics can be expanded through the concept of a "diversity-aware robot", i.e., an embodied AI that takes into account the unique characteristics of the person it is interacting with to be more easily accepted and provide them with better services. Examples related to research in different areas (including our own) are shown, ranging from "generative models" capable of producing images or gestures that are more likely to have a positive impact on people depending on what we know about their tastes to solutions for affective interaction capable to guess people's emotions and use this information to ensure them a more rewarding experience.