World Recipes Atlas is a culinary journey across the 196 countries of our planet—without ever stepping outside your living room. It is the first cooking book entirely generated by Artificial Intelligence, but imagined, questioned, and curated by two humans: Enrico Porfido and Juan Gonzalez del Cerro.
At its core, the atlas is a provocative artistic project, a reflection on how AI perceives gastronomic heritage and how it represents it. This book is a broader experiment. Not just in gastronomy, but also in technology and photography. What began as a playful challenge—asking ChatGPT to name and describe the most traditional recipe of every country—quickly evolved into a rigorous artistic and editorial process, questioning what AI knows, misunderstands, or invents about food traditions around the world, and wondering what are the differences between an image generated by AI and a photograph.
We did not set out to make a perfect cookbook. Instead, we embraced the imperfections, the surprises, and sometimes the absurdities of working with generative technologies. We approached the machine not as a tool, but as a co-creator—sometimes brilliant, sometimes clumsy, always revealing. We learned that while AI can mimic knowledge, it does not understand context, and that its mistakes can tell us as much as its successes.
This project is also a response to a growing need: to test and question the credibility of information produced by AI, especially in culturally sensitive domains. Through this book, we ask: what happens when culinary tradition is filtered through algorithms? What does AI prioritize? What does it miss?
