The Perfecto Project: Storytelling, Race, and Recognition in Human-Computer Interaction

This story is part of a series written by CCSRE's Public Writing Fellows.

 

Perfecto Flores is a middle-aged Mexican-American carpenter who leads a prototypical suburban life. He's happily married to Alicia, and they have two grown-up children, José, who is a single father, and Verónica, who recently got engaged. To all appearances, Perfecto’s life is simple and ordinary, almost perfect. But the truth is that the threat of a health catastrophe looms large on the horizon. Perfecto is pre-diabetic, and his doctor cautioned him against the perils of a sedentary lifestyle and unhealthy eating habits—so he decides to embark on a journey to a healthier life. And just like that, Perfecto's life becomes full of temptations that try to lead him astray from his goal.

What we have here are some basic building blocks for a good story—a relatable character, a quest, conflict, suspense, and so on—or so narrative theory tells us. Yet both literary critics and fiction aficionados alike will likely find this structural approach to storytelling insufficient. Something else might be missing.

So, what makes a story good?

This is the question that led James Landay, professor of computer science and associate director of Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI), to seek advice from Paula Moya, professor of English at Stanford. Landay and his team are currently working on a narrative-based fitness app that employs storytelling to motivate users to change their behavior to achieve their fitness goals. It all began when Landay and his team noticed that traditional fitness apps rely heavily on charts, graphs, and stats, which feel impersonal and often fail to keep users engaged in their workout routines. Worse still, in some cases, this antiseptic quantitative approach fuels negative attitudes towards physical activity, harming users' self-integrity, Landay and his team report.

So they created Zuki, a purple triangular-shaped alien from Thern, a planet recently ravaged by runaway global warming. Zuki's brother, who traveled to Earth to search for DNA samples from animals and plants to reignite life on Thern, gets into trouble and Zuki sets off in his rescue. WhoIsZuki is the fitness app that lets users follow Zuki’s quest if, and only if, they meet their exercise goals.

Moya joined Landay's team to offer advice on how to improve Zuki's narrative. But Moya, who was appointed Director of Stanford’s Center for Comparative Studies in Race & Ethnicity (CCSRE) last year, is more than a narrative expert. Her work with literature has long been driven by a more critical and politically urgent question—how literature reflects, promotes, and contests pervasive socio-cultural ideas about race, ethnicity, gender, and sexuality. As a literary scholar, Moya has developed critical reading skills that attend more carefully to the political dynamics underwriting literary representations in the stories she reads. This led her to propose expanding the WhoIsZuki project to include cultural-specific narratives that would be suited to motivating neglected populations, such as Latinxs or Native Americans.

The referenced media source is missing and needs to be re-embedded.

And that’s how Perfecto was born. With the assistance of Helena María Viramontes, a renowned Chicana fiction writer and professor of English at Cornell, whose fictional work has been of interest to Moya for some time, a group of interdisciplinary Latinx scholars and artists created a 13 chapter narrative for the app that features a traditional Latinx family in the U.S. Additionally, plans are underway to assemble a Native American crew for a new version of the Perfecto Project. Besides the well-known empathy argument, for Moya, part of what makes a story good is its capacity to let readers or listeners recognize familiar elements that make them feel valued, she explains in Matt Abrahams's Think Fast, Talk Smart podcast. The assumption behind the Perfecto Project is thus that recognition matters for good storytelling.

If raising questions about race, ethnicity, and other identitarian markers has changed the literary landscape in the US (and probably the world) in recent decades, it has begun to raise the stakes of the emerging field of Human-Computer Interaction (HCI). Race is a rather recent problem for HCI research—while most of the race-focused HCI papers were published after 2000, only a few of them published in the past few years have engaged seriously with race theory, exploring how conventional design practices perpetuate forms of both institutional and non-institutional racism. Nevertheless, these early explorations have already pointed to a crucial difference between literature and HCI, and it has to do with behavior.

User behavior is the nitty-gritty of HCI inquiry. For research like the Perfecto Project, obtaining hard data about the direct effects of culturally and racially diverse digital representations on people’s fitness behavior is key. The increasing improvement of digital and virtual technologies has enabled researchers to conduct tests that further complicate our views on the relationship between fictional (virtual or digital) representation and human behavior. The lessons learned on this subject are commonly known in HCI as the “proteus effect”—the idea that an individual’s behavior will conform to their digital or virtual self-representation, independently of how others perceive them.

Virtual reality (VR) has been particularly helpful in exploring the impact of the proteus effect due to the possibilities it offers for embodiment. For example, research has shown that users who use avatars with larger bodies behave more confidently, even more aggressively, towards their smaller virtual peers. Unfortunately, not enough experiments focused on race or ethnicity have been conducted to provide conclusive results. While some studies claim that white people embodying black virtual avatars can reduce their implicit racial bias, other similar studies have shown that VR embodiment can activate stereotypes in the presence of relevant features, such as gender or race. Researchers have tried to explain racial and gender stereotype activation through context, arguing that some situations are more triggering than others, for instance, job interviews for race and learning sports for gender.

People who are familiar with race theory or gender theory would have no difficulty in seeing the problem in this explanation. Markers like race and gender are two constitutive parts of our daily experiences as humans that cannot be just turned on and off at will. Despite the lack of scientific evidence of their biological existence, race and gender are key social characteristics with real impact on the world; these markers make people intelligible to each other and to themselves. Virtual environments refer outward inevitably to the real world, and finding real-life situations where race or gender are unimportant is simply not possible.

And that's why race-informed storytelling is important for HCI. Understanding why we love good stories can allow us to become aware of fallacious, deep-rooted assumptions about reality that permeate all aspects of our practices—for example, the presumption of neutrality. Originally, the idea to have a purple alien starring the story in the fitness app was to be as neutral as possible. In explaining her motivations to join the WhoIsZuki team, Moya said that Zuki, the seemingly genderless alien, struck her as male. After all, Zuki was on an epic journey to save the planet, similarly to the epics where male heroes have always embarked on in our tradition to save the world. The hero's journey is compulsorily a male-dominated literary structure.

If every situation in the metaverse will unavoidably trigger people's mental schemas about race and gender, then perhaps one way to fight against them is through heightened exposure. That is, perhaps by increasing diversity in racial, ethnic, and gender representations in the stories we tell we can overexpose people until the triggers wear off. In this respect, the Perfecto Project might be on the right track.

Asking questions about race, gender and cultural differences has taught Moya that stories are never neutral, they come with baggage. And now, those questions are teaching us that designing technology might not be so different from crafting a good story after all. Telling Perfecto's story is as much a way to combat color-blind tendencies in tech design as to include overlooked user populations. As Moya says, representation matters in the way we craft both the stories we tell and the technology we use.

Learn more about The Perfecto Project here.

Alberto Quintero is a 2022 CCSRE Public Writing Fellow and a Ph.D. Candidate in Modern Thought & Literature at Stanford University.