The Unboxing_AI: an introduction workshop took place on May 21, in an online format and was attended by faculty members, researchers, PhD students, and other DigiMedia collaborators.

Joaquim Santos, an invited Assistant Professor in the Communication and Art Department, and researcher at DigiMedia, led the workshop, which main goal was to provide an accessible yet informed introduction to Large Language Models (LLMs), covering their evolution up to the Transformer architecture and the Generative Pre-Trained (GPT) models. The workshop explored codeless prompting techniques and the use of LLMs in academic research, including their limitations, risks, and the current regulatory framework.

The discussion also included the use and deployment of local LLMs versus cloud-based LLMs, weighing their advantages and disadvantages. A set of useful resources was presented to facilitate more effective interaction with AI systems. 

In the hands-on session was an opportunity for participants to build knowledge using Open WebUI, integrate it into a Retrieval-Augmented Generation (RAG) system, and create a functional chatbot based on that knowledge.