Carlos Miguel Ríos-González · American Journal of Sexuality Education 2026 · 2026
DOI: 10.1080/15546128.2026.2742456
Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).
The convergence of artificial intelligence (AI) and human sexuality is reshaping intimacy, sexuality education, entertainment, and therapeutic practice. This narrative review analyzes these transformations through the complementary lenses of Sexual Script Theory (SST; Simon & Gagnon, Citation1986) and technological mediation theory (Verbeek, Citation2012), which together illuminate how AI does not merely respond to preexisting sexual behaviors but can mediate, generate, and transform sexual scripts at individual, interpersonal, and societal levels. For the purposes of this review, AI is defined broadly to encompass both traditional machine-learning (ML) systems—such as recommendation, ranking, and predictive algorithms—and contemporary generative AI (GenAI), including large language models (LLMs) such as ChatGPT and Claude and multimodal systems capable of generating synthetic text, images, audio, and video. The review examines five domains selected for their prominence in the recent literature and their capacity to illustrate different levels of script formation and technological mediation: sex robots and AI companions, AI-generated pornography, algorithmic dating and AI-mediated romantic communication, AI-assisted sexuality education, and AI-supported sexual health and therapeutic applications. Across these domains, AI offers opportunities to expand access to information, connection, and support while simultaneously raising concerns about consent, privacy, objectification, bias, authenticity, and cultural representation. An ethically responsible approach therefore requires attention not only to what AI systems produce, but also to the sexual scripts, relational expectations, and values that their design and use may normalize.
No comments yet — start the discussion below.