An analytical framework for studying attitude towards emotional AI: The three-pronged approach

Abstract

Emotional artificial intelligence (AI) is a narrow, weak form of an AI system that reads, classifies, and interacts with human emotions. This form of smart technology has become an integral layer of our digital and physical infrastructures and will radically transform how we live, learn, and work. Not only will emotional AI provide numerous benefits (i.e., increased attention and awareness, optimized productivity, stress management, etc.), but in sensing and interacting with our intimate emotions, it seeks to surreptitiously modify human behaviors. This study proposes to bring together the Technological Acceptance Model (TAM) and the Moral Foundation Theory to study determinants of emotional AI's acceptance under the analytical framework of the Three-pronged Approach (Contexts, Variables, and Statistical models). We argue that to quantitatively study the acceptance of new technologies, it is necessary to leverage two intuitions. The first is the degree of acceptance increases with how users of smart technology perceive its utilities and ease of use (formalized in the TAM). The second is the degree of acceptance decreases with the user's perception of threat or affirmation posed by the technology in relation to social norms and values (formalized in the Moral Foundation Theory). This study begins by mapping the ecology of current emotional AI use in various contexts such as workplace, education, healthcare, personal assistance, etc. It then provides a brief review and critique of current applications of the TAM and the Moral Foundation Theory in studying how humans judge smart technologies. Finally, we propose the Three-pronged Analytical Framework, offering recommendations on how future studies of technological acceptance could be conducted from the questionnaire design to building statistical models.