These days, we are surrounded by creative text generators like GPT that seem to be able to generate any text we want. But how do we ensure that AI truly aids us in overcoming our unique creative challenges? In this talk, we uncover the remarkable potential of "prompt engineering" – the art of enhancing our communication with AI. We dive into the world of autoregressive text generators, learn their inner mechanisms and which training phases they went through to get to the current state-of-the-art text generators. These insights help understand why certain prompt engineering techniques (such as few-shot prompting, role-prompting and chain-of-thought prompting) are able to outperform simpler prompting methods. We show how even some of AI's classic hard problems, such as humor generation, become even more within reach thanks to these large language models and their prompt engineering techniques.

Thomas Winters is a researcher in creative artificial intelligence and computational humor at KU Leuven. In his research, he develops neuro-symbolic AI models for aiding in co-creative humor writing tasks.