AI-Generated Menus Trigger Uncanny-Valley Backlash

Restaurants are turning to generative tools such as ChatGPT and Midjourney to produce menu art without repeated photo shoots. Yet models learn visual patterns rather than the physical nature of food, often averaging familiar advertising styles from chains including McDonald’s, Burger King and Wendy’s. The result — flawless symmetry, rounded shapes and unnaturally smooth textures — can fall into an “uncanny valley,” eroding appetite and trust when diners cannot tell whether an image represents a dish the kitchen actually serves.

TechCrunch reported on Sept. 3, 2026, that the effect can intensify through repeated edits. An X user named Labtec posted an Aug. 19 experiment in which a ChatGPT-generated menu was edited 100 times, with the food becoming progressively rounder and less realistic. Reality Defender Chief Technology Officer Alex Lisle described the pattern as convergence, not full model collapse. Research from Germany’s University of Duisburg-Essen also found that near-real AI food images prompted more disgust and unease than obviously fake ones, strengthening demand for AI-detection and content-verification tools.