Prof. Emmanuele CHERSONI, Department of Chinese and Bilingual Studies

Large Language Models in Humanities and Arts. Seminar organised by CEFC and NTU. Centre d'études français sur la Chine Contemporaine (CEFC), Graduate Institute of Linguistics, National Taiwan University, Taipei, Taiwan, 3 March 2025.

Abstract
Since the introduction of the ChatGPT conversational chatbot in November 2022, Large Language Models (LLMs) have been at the forefront of the current AI revolution. Their strong performance on NLP benchmarks, together with their striking naturalness in conversation, contributed to stimulate a new debate in public discourse and media about the imminent arrival of Artificial General Intelligence (AGI). It has been shown that LLMs can display new abilities and generalizations that were not predictable simply on the basis of parameter scaling-up (i.e. emergent abilities), which also prompted the question on whether such systems can generate novel concepts and ideas beyond what they have seen during their training – or, in other words, whether they can be creative.

Although machines may not be creative in the same way as humans are, the fact that AI-generated texts and images are becoming more and more undistinguishable from human creations raised a lot of debate and the understandable fear that, consequently to the application of the new technology in fields such as journalism, literature and visual arts, many jobs are going to disappear.

In my talk, I will try to address the following question: What are the applications of LLMs to the Humanities, and what is next for our disciplines? I will focus on a specific case study from the field of Chinese Digital Humanities: the usage of LLMs for automatizing the translation of the Black Myth Wukong videogame.