While recognising the importance of artificial assistance systems, the project studies their involvement during the training process. This is done from the perspective of human learning (cognitive sciences), machine learning (computer sciences), and by analysing the trust in AI partners (philosophy).
Digital touch sensors are increasingly coupled with artificial intelligence to assist humans in their daily and professional activities, for example, in lane departure warning systems or when robots assist with precision surgery. The question this project aims to answer is: should we train with or without this artificial helping hand? Cooperative learning, which occurs when two agents aim to learn certain attributes together, is likely to involve not just human peers, but hybrid pairing of human and artificial learners. Using a novel interdisciplinary approach, this project team examines human-AI hybrid learning for increasingly innovative tactile augmentation and assistance by integrating three different, but complementary perspectives: the cognitive neuroscience of human, biological learning through vision and touch, the philosophy of self-confidence and trust in digital tactile assistants and computer science design of machine learning algorithms tailored to tactile learning with AI. The project includes citizen-science components from medicine and driving practice, which will help to transfer the results to a practical application. Together, these timely initiatives will pave the way for the introduction of this new technology into society and will increase the end user’s willingness to make the best use of these assistance systems.