A technology leader perishes in an autonomous vehicle operated by a self-driving company. A humanoid AI apparatus acts as the family caretaker for an elderly parent. These themes were examined in two original plays crafted and showcased at MACH 33: The Caltech Festival of New Science-Driven Works, which focused on artificial intelligence during its 2025 season.
Now in its 13th iteration, MACH 33 connects independent writers with scientific consultants from Caltech and NASA’s JPL, overseen by Caltech, to create plays rooted in science or technology, enriched by specialized knowledge. The Null Test, penned by Ashley Quach and guided by Caltech Professor of Philosophy Frederick Eberhardt, revolves around a tragic incident involving an autonomous vehicle and the subsequent legal proceedings, in which the liability—and agency—of the algorithm is scrutinized. “It was astute of MACH 33 to pair me with an ethicist,” Quach states. “You receive vastly different insights on AI when conversing with an engineer or a data scientist. A philosopher offered the ideal equilibrium for addressing the dilemmas I investigated in my play.”
Dementia Doula, authored by Tom Lavagnino and guided by Joel Burdick, Caltech’s Richard L. and Dorothy M. Hayman Professor of Mechanical Engineering and Bioengineering and a research scientist at JPL alongside his graduate assistant Solvin Sigurdson, envisions a situation wherein an AI caregiving robot integrates into a family’s routine as they confront a relative’s aging and memory decline. While Lavagnino admits he drew motivation for the play from his own experience seeking a caregiver for his father diagnosed with Lewy body dementia, Burdick and Sigurdson contributed to the realistic depiction of the embodied AI robot.
“We deliberated on what a robot can feasibly accomplish today or in the near future so that when audiences view the performance, they engage more with scientific reality rather than science fiction,” Sigurdson explains. Nonetheless, the scientific advisors also warn that some representations in the play—such as a robot’s capacity to perform diverse tasks efficiently, or “generalizability”—remain significant hurdles within the field.
Each of the two works was delivered as a staged reading featuring an ensemble cast of Caltech scholars, community participants, and actors from Los Angeles. Furthermore, the festival incorporated “Storytelling for the Future: A Symposium on AI, Ethics, and Playwriting,” which included a panel discussion among the playwrights and scientific advisors, moderated by Travis Cloyd, a producer in residence at CalArts and CEO of media firm Worldwide XR. During the symposium, the panel members shared their collaborative experiences and explored scientific and creative concepts.
Eberhardt noted that when he was initially invited to serve as a scientific advisor, he envisioned something akin to “correcting the physics of the narrative,” like elucidating how gravity might function on a flat planet. “But that wasn’t my role as a scientific advisor at all,” he remarked. “The conversations I had with Ashley were far more profound, focusing on the science and ethical consequences. We discussed notions to clarify the underlying considerations pertinent to her narrative, such as What role does AI serve? What constitutes agency for artificial intelligence? And at what point do we contemplate a moral standing for an AI?”
The panelists engaged in deeper inquiries regarding similar ethical dilemmas during the symposium, expressing diverse viewpoints and levels of optimism pertaining to AI’s future. Quach voiced her apprehensions about the dangers linked to AI advancement, particularly in situations where a limited number of influential individuals make significant decisions with minimal oversight.
“There’s this way of discussing progress in the tech sector that presents it as if it’s gravity, an unavoidable force of nature,” Quach remarked. “For me, [AI advancement] isn’t fundamentally a moral or ethical issue; it’s primarily about deceleration. Can we please engage in dialogues regarding the risks? Can we practice responsibility towards each other?”
Burdick and Sigurdson highlighted the prospective advantages of AI and robotics, especially within the healthcare domain, as portrayed in Lavagnino’s play. Burdick mentioned that his research experience in developing robotic exoskeletons designed to assist patients with spinal cord injuries provided him insights into how technology can be beneficial as well as into the regulatory framework.
“I had to present in front of the FDA panel, and the procedure was simpler than one might anticipate,” Burdick explained. “They seek to determine what risks exist for the patients and what the benefits or rewards are for them. Regarding AI, we are progressively learning about the hazards, but I don’t believe we completely grasp them at this moment. The question that preoccupies me is, How can we ascertain those risks and benefits in a timely manner and in a way that minimizes potential harm while emphasizing the advantages?”
Eberhardt suggested that the optimal way to “integrate ethics into AI” would entail contemplating the ultimate intended use of a technology. “I don’t think you establish [ethics] from the ground up,” he explained. “The context and implementation of the AI or algorithm will mold the ethical and moral considerations. However, that becomes complex when the tools are increasingly generalized and not developed for a specific application.”
The panelists concurred that even with the transformative benefits AI can offer in certain sectors, regulating the technology is crucial to lessen risks and safeguard human health, welfare, and safety. An additional danger they discussed was the risk of individuals becoming overly reliant on AI and forfeiting crucial skills—which led to a discussion regarding the essence of creativity.
“AI can be an exceptional instrument, but I believe it can also entice you into losing some foundational abilities that may be vital,” Sigurdson observed. “Depending on what you’re attempting to accomplish, you need to critically evaluate whether you should depend on this tool. Especially as a researcher, it’s essential you understand your foundations; for without that knowledge, how can you innovate and develop something entirely original?”
Quach recounted her experience crafting The Null Test: “There’s something truly enchanting in permitting yourself to err, to embrace the poor ideas and to compose the drafts that aren’t up to standard. Because this process strengthens you; you improve with every endeavor.”
The playwrights also pondered whether AI-generated art could possess significance for humans. “Can an AI narrative ever match the authenticity and fulfilling experience of reading one composed by a human being?” Lavagnino queried. “I believe the verdict is still pending. It will be intriguing to see how it unfolds. I mean, I’ve certainly never encountered an AI-generated tale that could measure up to the works of human authors thus far.”
“I think it genuinely relies on the audience,” Quach reflected. “Ultimately, you, as the reader, are the one who judges the worth. If you find significance in a narrative you’ve engaged with, I cannot contest that. It’s your personal connection to the piece; if it offers you something meaningful, that’s valid.”
She continued: “But for the writer, it’s distinct, for what is the motive behind crafting a narrative beyond mere compulsion? I don’t even possess an answer to this myself. Why? What is the intrinsic value? What is the objective? I know that during the writing journey, it resembles breaking open a piñata. You approach a topic unaware of what you’ll uncover. You’re simply gathering fragments of yourself, little treasures from your memory and subconscious, which allows you to engage with yourself in a novel manner. It enables you to heal, express, and connect with those you cherish. For me, that’s an extraordinarily noble pursuit.”
MACH 33 is directed by artistic lead Brian Brophy, who oversees Theater Arts at Caltech (TACIT); associate artistic director Arden Thomas; and Cole Remmen, who serves as the program director for Launchpad, an early-development play lab working within MACH 33. The MACH 33 festival is made possible through generous contributions to Theater Arts at Caltech.