Human-Robot Interaction Fostered by Generative Artificial Intelligence

Assistive Social Robotics (SARs) refers to robots that assist human users. With technological advances and studies in Generative Artificial Intelligence, new research and projects in SARs are being developed. The OpenTera framework currently does not encompass the potential of these technologies to facilitate SARs' development. Additionally, it does not include specific conversation modules, which are essential for the robot's social functions. The objective of this study is to make adaptations to the OpenTera framework to incorporate the benefits of using the ChatGPT API in SARs' development. We also take into analysis the privacy threat involved in the design of such a prototype. The study and development were carried out by a mobile application in Flutter, using libraries for automatic speech recognition and speech production, which communicates with the ChatGPT API to respond to user queries. The study's results led to the development of a new module for adapting the OpenTera framework. Considering the results, the new module can be added to the framework, making it more adequate, with an integrated SARs module powered by generative artificial intelligence. This is an interdisciplinary and original study between Law and Computer Science, in which the results show that privacy and personal data are vulnerable to the use of new technologies.