On February 3, Nina Beguš, a researcher and lecturer at UC Berkeley, presented her work "Artificial Humanities" to attendees at the BIDS Suite in Sutardja Dai Hall and online. We in the audience were invited to understand how fictional narratives shape the human creation and understanding of artificial intelligence.
Beguš focused on George Bernard Shaw’s Pygmalion: published in 1911 and popularised on the silver screen in 1938, and through the Broadway musical My Fair Lady in 1956. In the story a man — Henry Higgins — takes on a challenge to fool high society into thinking a London flower girl — Eliza Doolittle — was a member of the upper class. Nina linked Eliza’s story to the Turing Test and to the ELIZA chatbot.
Data scientists and computer scientists have likely heard of the Turing Test: can a computer trick a person into labelling its side of a conversation as human generated? The ELIZA chatbot extends this question: Do people conversing with a chatbot attribute human-like feelings and reasoning to the computer programme?
The Turing Test was proposed in 1950 and the ELIZA chatbot published in 1966. Assigning inanimate objects human attributes is not unique to large language models or robotics. (In fact, as Nina shared, the title of Shaw’s story goes all the way back to Greek myth: Pygmalion was a sculptor who fell in love with a statue he had carved.) However, these patterns continue to persist in our fictional and scientific narratives. Beguš drew connections from recent popular movies, such as Her and Ex Machina, to current virtual assistant technologies, virtual therapists, and robotic companions.
A slide from Beguš’s talk. There are four images in a 2x2 grid with a title of “fictional and actual Elizas’. Top left is an image of the social robot Ava from Garland’s film “Ex Machina” (2014), top right is the social robot Sophia of Hanson Robotics (2018). Bottom left is Eliza Doolittle from the film Pygmalion (1938) and bottom right is a screenshot of Joseph Weizenbaum’s chatbot ELIZA (1966).
One of Beguš’s recent publications: “Experimental narratives: A comparison of human crowdsourced storytelling and AI storytelling” (Beguš, 2024) asked both human crowdworkers and large language models to write short stories about creating and falling in love with an artificial human. Both humans and machines perpetuated the Pygmalion myth, meaning both humans and machines created stories about someone falling in love with a human-like creation.
In her work Beguš asks “What do humanities provide to technology?” Her answers:
- Analyze historical and current conditions together with ideas about the future
- Address cultural aspects of AI development
- Fiction offers a public space for ideas, speculations, ramifications
- Vocabulary, articulation of the new
- Innovative thinking not tied to a stable ontology
Importantly, Beguš reminded us not to limit our imagination to building machines that can emulate humans. Where can machines extend, or augment, human creativity? How can we use our most human characteristics of creativity in language and storytelling to envision exciting and valuable uses of technology?
Watch the video below and access the presentation slides at https://osf.io/9p3ht/!
Read more of Nina Beguš’s work:
- Beguš. (2024). Experimental narratives: A comparison of human crowdsourced storytelling and AI storytelling. In Humanities and Social Sciences Communications (Vol. 11, Issue 1). Springer Science and Business Media LLC. https://doi.org/10.1057/s41599-024-03868-8
- With data openly available at Beguš. (2025). Data for Experimental Narratives: A Comparison of Human Crowdsourced Storytelling and AI Storytelling. https://doi.org/10.17605/OSF.IO/K6FH7
- Pre-order her book “Artificial Humanities” at https://doi.org/10.3998/mpub.12778936
