Picasso and AI is an educational initiative designed to enhance the understanding of artificial intelligence models among students of art history, visual arts, and museum studies (https://catedratelefonicauma.es/actividades/picasso-y-la-ia-un-taller-para-analizar-y-cuestionar-la-tecnologia-que-nos-desborda/). The primary goal is to guide participants through understanding the operational logic of AI models by engaging in the design and execution of a specific exhibition project. Critical reflection on issues inherent in these models, such as biases, is an integral part of the workshop. The project takes Picasso’s extensive artistic production as a starting point for exploring the critical issues embedded in AI models, as well as the potential opportunities for enhancing curatorial narratives. Using a customized corpus of both Picasso’s visual works and writings about Picasso from recent decades, participants employ various AI models, in particular chatGPT and CLIP by OpenAI, to craft an exhibition narrative about the artist and object that would be displayed (https://github.com/BarbaraRomero/Picasso-AI). The methodology used for the creation of this workshop involves collecting bibliographic references and abstracts about Picasso, categorized by decade, along with images of his artistic output. The workshop unfolds in several phases, wherein participants use generative AI tools to develop the textual content and narrative of their exhibition based on the collected bibliographic records. Once the exhibition narrative is decided, participants use an online platform built ad hoc for the project (http://amozo.es/PicassoAIg/). This platform has been developed using the CLIP model to source visual resources among Picasso’s artistic production for their exhibition so that participants can explore Picasso’s visual production through the concepts involved in the generative curatorial narrative. Each group presented their generated narrative and selected Picasso’s works, as well as discusses the issues and challenges that arouse during the process. The workshop aims to shed light on the inherent biases and limitations present within these AI models. Participants gain insights into the utility of these tools and learn how to leverage them effectively while acknowledging their imperfections. This understanding is critical to preventing overreliance in educational settings and enables participants to navigate and interact with generative AI tools.