“AI Unleashes Ancient Greek Drama: Professor Completes Sophocles’ Lost Play – Can You Spot the Difference?”
In the city of Johnstown, Pennsylvania, the renowned play “The Searchers” by the ancient Greek playwright Sophocles has remained incomplete for over two millennia, with only the first three acts recovered. Mark DiMauro, an assistant professor of digital humanities at the University of Pittsburgh at Johnstown, recently utilized artificial intelligence to “complete” the work.
On November 30th at 5 p.m., portions of the play will be performed by Pitt-Johnstown theater students in Room 131 of Blackington Hall. DiMauro explained that the group will read from both the original play and his AI-generated additions, inviting the audience to discern any differences and fostering a discussion on the applications of AI.
Sophocles
Having worked with artificial intelligence for approximately five years, DiMauro explored its creative potential as the tools gained popularity. Intrigued by the idea of using AI for incomplete works that might never be finished, he turned his attention to “The Searchers” due to the availability of preserved content, even though certain sections of the play are likely lost forever.
The play revolves around the newborn god Hermes, who plays a prank on his older brother Apollo by stealing his cattle. Apollo then dispatches a group of satyrs to recover the herd, promising substantial rewards.
To undertake this project, DiMauro initially used a series of JavaScript data agents before transitioning to ChatGPT 4, a large language model-based chatbot developed by OpenAI. Applying distant reading, which involves understanding literature by analyzing vast amounts of data rather than examining specific texts, DiMauro added sentence structures, word choices, and related information to reconstruct missing parts of the play.
DiMauro emphasized that the reconstruction is as close as possible to Sophocles, with the need for further examination by others to validate the program’s accuracy.
The project received support through a Pitt Momentum Funds micro-grant, with contributions from Pitt-Johnstown professors Jeremy Justus, Derek Leben, and Ken Sherwood.
While excited about the potential applications of AI, both DiMauro and John Teacher, an assistant professor of theater arts at UPJ, expressed concerns about the technology being used to replace humans or for plagiarism. They view AI as a valuable tool for gaining insights into historical texts and believe that the humanities will play a crucial role in guiding the ethical use of this technology.
Michael Stoneham, UPJ humanities chairman and associate professor of English literature, praised DiMauro’s project as a pioneering use of generative software for textual restoration, describing it as an opportunity to contribute to the understanding of various fragmentary texts.
The collaboration between DiMauro and Teacher aims to bring a dialogue on the use of artificial intelligence in the humanities, particularly in the restoration or reconstruction of incomplete ancient works, to the campus, delving into the intricacies of this fascinating subject.
