This chapter critically examines AI-driven recipe generation, from early efforts like IBM’s Chef Watson to current tools such as Chef-GPT. While AI has advanced in generating and adapting recipes, it continues to fall short of creating genuinely original or satisfying culinary innovations. The limitations of these systems reveal the complexity of modelling multisensory flavour perception, which depends on the dynamic interplay of taste, aroma, texture, and visual appeal, shaped by culture, genetics, and individual experience. The transformations that ingredients undergo during preparation further complicate computational predictions. Although AI has been applied in commercial product development, for example by McCormick’s and Coca-Cola, attempts to digitise the chemical senses have yet to succeed. The chapter argues that, while AI may assist in refining existing recipes, it struggles to replicate the creativity, intuition, and contextual sensitivity that define human cooking. These challenges underscore the gap between algorithmic outputs and lived sensory experience. However, advances in sensory science and computational modelling may gradually narrow this gap in the years ahead.