ELIZA: Finding love & human connections through algorithms

The lights dimmed, and onstage sat a humble collection of props: a wooden bench, a couple of books, a broomstick and dustpan. Unfolding amidst this innocuous mise-en-scène, Eliza – written by Professor José Ignacio Latorre (Director, Centre for Quantum Technologies) – sought to explore how relationships between man and robot could unravel, and with that, the central premise of the production: Is artificial intelligence (AI) capable of love?

The play tells the story of Adam, a lonely man seeking companionship through a series of robot helpers, each becoming increasingly human-like. Rejecting earlier versions, Adam finally falls in love with Désirée, the third prototype, who seemingly reciprocates Adam’s confessions and advances. As Adam progresses to Eliza, the next – and final – iteration of his artificial companion, the play leads audiences to uncomfortable revelations about Adam and Eliza, ending on a cliffhanger that takes audiences back to the play’s central premise. While the play, by design, did not provide specific narrative closure, it certainly led audiences to question what constitutes artificiality and intelligence, as well as why and how people express love.

A production with such lofty philosophical aspirations could have easily fallen flat, but was adeptly brought to life through spellbinding performances by lead actors Jayesh Melvani (24, Faculty of Law) and Tinotenda Zimhunga (20, Yale-NUS College), respectively playing Adam and a series of robot helpers, including the titular character Eliza.

Tinotenda’s dynamic and disciplined delivery was undeniably a highlight of the production. She transitioned seamlessly through a multitude of AI “prototype” personae, each one becoming more complex and technologically advanced than the last, with their own unique personality and mannerisms. In counterpoint to Tinotenda’s character, Jayesh’s portrayal of the human Adam provided a relatable point of reference that complemented Tinotenda’s shifts in character, as he ranged from seeking intimacy, to being intimidated by a superior being, and to falling in love.

In a post-performance dialogue session, Professor Latorre shared that the play was inspired by a Massachusetts Institute of Technology-created AI algorithm codenamed “Eliza”, which was designed to be able to relate emotionally to humans. He observed that people were attributing human-like emotions to AIs, as humans seemed increasingly robot-like, remarking, “Human predictability might pave the path to programming advanced AI that will provide us with company, given the right balance of care, confrontation and friendship”.

It is this fundamental tension between the nature of humanity, and the weight that humankind places on appearances, visual and verbal cues in distinguishing between organic and artificial, that drives the play’s premise and successfully challenges the audience’s ideas of humanity, AI, and emotion. “If you can't distinguish between a machine and a human when exchanging messages, then we have to view that algorithm as intelligent,” claimed Eliza during the play, ”What matters, Adam, is not being real, but being indistinguishable from something real."

Presented by NUS Centre For the Arts as part of the ExxonMobil Campus Concerts series, Eliza was directed by Associate Professor Maria Teresa Soto-Sanfiel (NUS Department of Communications & New Media, and Principal Researcher at the NUS Centre for Trusted Internet and Community) and was staged live from 24 to 26 February at University Cultural Centre Theatre. You can find out more about ExxonMobil Campus Concerts here.

By Nur Azzahra Binte Kamal, NUS Centre for the Arts