Follow the latest coverage, related explainers and connected technology stories.
The article explains how training data and “satisfaction” criteria push artificial intelligence models to produce smooth, symmetrical food images that lack realistic details. The problem is not limited to restaurants, as it reveals broader risks associated with the convergence of model outputs and the erosion of trust in visual content.