03 - From the Presocratics to ChatGPT: Teaching Classics and the Ethics of AI

Artificial intelligence has become an integrated element of our daily lives, transforming our research and classrooms. This was already evident in 2020, when I first taught a course in Classics and AI. Beginning with Heraclitus and OpenAI, and focusing on the ethical complications then delaying the release of what would become ChatGPT, the course went on to introduce students to a wide range of ancient sources from the Graeco-Roman world, familiarizing them with modern scholarship in ancient science and philosophy (Mayor 2019), and facilitating debate about issues such as - whether artificial moral agents can exist (Van Wynsberghe 2019); why we want to humanize AI and whether we should (Watson 2019); whether intimacy can exist between humans and AIs (Fiske et al. 2019); and what the implications of AI are to human intellectual, creative, and empathic capabilities (Verganti et al. 2020).

Based on my experience teaching the course, I believe courses like it have a place in all Classics departments. Classical material is effective in confronting students with difficult questions pertaining to the development of AI, our engagement with technology, and the historical-psychological processes that have brought us to the precipice upon which we now stand. Structuring it around familiar elements of AI technology and creating assignments in which students can engage with AI platforms provides excellent cognitive scaffolding for iGen students. I witnessed the high degree of dexterity students have with concepts that are native to their experience, and how this familiarity helped to engage those students who might otherwise find ancient philosophy and poetry difficult to access. Consequently, I believe such a course can benefit all students, Classics majors and non-majors alike.

Still, for Classics majors, an AI-centered course bolsters their ability to respond to the advancements that AI will bring to the field in the near if not immediate future. Large language models can already summarize what would take days or weeks to read; identifying authentic works and filling in lacunae are on their way to becoming fully automated tasks and machines soon will be able to imitate any ancient author and produce "new" works. The knowledge we seek as classicists will necessarily change with the technology. It is thus imperative that students be prepared in two ways. First, they must be skilled in assessing the output of the machines, evaluating their claims, and reasoning about the attention-shaping effects of technology. Also, they must be prepared to respond to the changes that are coming to the field as a result of the previously unimaginable advances that AI makes possible. The humanities will surge in value as specialized humanistic knowledge is increasingly and conspicuously applied to social and scientific problems (see, e.g., Atari & Henrich 2023). Classics courses that examine ethical dimensions of AI represent a first step in preparing students and classical scholars to thrive on this new frontier, both guiding their engagement with AI tools and providing a foundation from which to engage with the challenges of our time (see, e.g., Burton et al. 2017).


Presenters

Jennifer Devereaux, Harvard University



  SCS-9