Interrogating Large Language Model Discourse on Pornography: Implications for Sexuality Education

Large Language Models have had a massive impact on learning, knowledge production, and educational processes. Concerns regarding the reliability of the knowledge LLMs produce, the impact of their use on critical thinking, and how to promote AI literacy, have been raised by many critical researchers, especially in relation to sensitive topics. This paper explores implications of the use of LLMs for comprehensive sexuality education and pornography literacy. Discourses generated by popular LLMs when asked about pornography are analyzed through a feminist lens. The LLMs produced power-blind statements about pornography, creating silences around the gendered practices. Moreover, the LLMs produce reduced discourses around pornography, violence against women, and consent. When asked to generate lesson plans, simplistic plans were created that lack systemic analysis of pornography and limit critical thinking to individual awareness. The paper argues that if LLMs are applied in CSE, the discourses they generate must be carefully considered through feminist and critical pedagogical perspectives that address power relations, oppression, and violence within the wider social context. Furthermore, the results suggest that the proliferation of easily accessible AI-generated knowledge makes comprehensive sexuality education curricula about pornography and teachers’ training even more essential.