Beyond “tasty but nasty”: AI-guided food pairing to re-imagine edible insects in gastronomy

Growing pressure on global food systems has intensified the search for sustainable, protein-rich alternative ingredients, positioning edible insects as promising yet culturally contested food sources, particularly in Western societies. This study proposes an interdisciplinary framework that integrates Food Pairing Theory with artificial intelligence-assisted recipe prototyping to reframe entomophagy through gastronomic compatibility rather than novelty or necessity. Adopting a qualitative single intrinsic case study design, the research operationalizes food-pairing logic by identifying insect-ingredient combinations with shared aromatic compounds and translating these pairings into structured recipe prototypes using a large language model (ChatGPT-5.2). Three edible insect categories – beetles, caterpillars, and bees – were selected based on their global prevalence and nutritional relevance and paired with complementary plant-based ingredients using a data-driven aroma-matching database. Standardized prompts were employed to generate draft recipes, which were subsequently refined through a systematic expert evaluation process. Two gastronomy-trained experts assessed the finalized prototypes using a structured 7-point anchored rubric addressing pairing coherence, anticipated sensory balance, culinary feasibility, clarity, gastronomic positioning, innovation, and food-safety awareness. Inter-rater reliability was examined using quadratic-weighted Cohen’s kappa ( κ w). Findings indicate that AI-assisted recipe prototyping can coherently translate food-pairing theory into gastronomically plausible concepts, with particularly strong performance in pairing coherence and novelty-with-plausibility dimensions. The study contributes a transparent, reproducible workflow for integrating artificial intelligence into gastronomic innovation and demonstrates how taste harmony can serve as a strategic lens for addressing consumer resistance to insect-based foods. Future research should extend this framework through sensory testing, consumer acceptance studies, and comparative evaluations across insect species and algorithmic pairing systems.