Enhancing the Trustworthiness of Generative Artificial Intelligence in Responsive Pedagogy in the Context of Humanities Higher Education

Abstract

How do we enhance the trustworthiness of generative artificial intelligence (AI), such as ChatGPT, as a tool to foster students’ curiosity to learn about humanities subjects in higher education? This study analyzes what conversational AI tools can realistically accomplish in the humanities higher education context and what the substantive, rather than hyped, challenges are. There are two challenges. The first challenge is false singularity. While the AI has deficient domain knowledge, it is able to simulate fluent prose which can be mistaken as the ultimate answer to a query. One solution is to promote critical AI literacy, which enables students to grasp the nature of AI-powered simulation, and to nurture metacognition, a self-reflexive understanding of one’s own learning and thought processes. The second challenge is the tendency to mistake AI synthesis for critical thinking, the solution to which is the flipped classroom. Instead of writing essays that respond to instructor-generate prompts, students construct open-ended but focused research questions that are refined through reiterative and interactive activities. The author has constructed her own AI Teaching Assistant to test these hypotheses.

Keywords: Generative AI, trustworthy AI, humanities, arts, higher education, critical questioning skills, meta-cognition, Shakespeare