AI-Powered Procedural Haptics for Narrative VR: A Systematic Literature Review

Haptic feedback is important for narrative virtual reality (VR), yet authoring remains costly and difficult to scale due to device-specific tuning, placement constraints, and the need for semantically congruent timing. We systematically reviewed user studies on haptics in narrative VR to establish an empirical baseline and identify gaps for AI-powered procedural haptics. Following PRISMA 2020, we searched IEEE Xplore, ACM Digital Library, Scopus, Web of Science, PubMed, and PsycINFO (English; human participants; haptics synchronized to narrative events) and performed backward/forward citation chasing (final search: 31 July 2025). We also conducted a parallel scoping scan of grey literature (arXiv and CHI/SIGGRAPH workshops/demos), finalized on 7 September 2025; these records are summarized separately and were not included in the evidence synthesis. Of 493 records screened, 26 full texts were assessed, and 10 studies were included. Quantitatively, presence improved in 6/8 studies that measured it and immersion improved in 3/3; sample sizes ranged 8–108. Across varied modalities and placements, haptics improved presence and immersion and often enhanced affect; validated measures of narrative comprehension were rare. None of the included studies evaluated AI-generated procedural haptics in user studies. We conclude by proposing a structured, three-phase research roadmap designed to bridge this critical gap, moving the field from theoretical promise to the empirical validation of intelligent systems capable of making rich, adaptive, and scalable haptic narratives a reality.