Prompt Design and UX principles to unravel the hallucinations of LMMs (by Federica Nardi, Federica Ramazzotti) In the last year we have heard more and more about Chat-GPT and the concept of "prompt". Compared to this scenario, the Prompt design stands as the art of conversing with generative AI to get the desired outputs. However, these do not always meet expectations. So the prompt design is becoming the bulwark to contain errors and "inventions" of these models, especially when it comes to written output. We are talking about hallucinations. What are hallucinations? Are they necessarily bad? In this regard, what are the fundamentals for a good prompt? And how do they intersect with UX principles? In this talk we will analyze the limits and possibilities of hallucinations in order to set a correct and/or creative output.