Providence Reporter

By merging themes in dance and computer science, the course Choreorobotics 0101 is teaching the next generation of engineers how to create technology that minimizes harm and makes a positive impact on society.

Choreorobotics 0101 is helping the next generation of engineers create emerging robotics and AI technologies that minimize harm and make a positive impact on society. All photos and videos: Nick Dentamaro/Brown University

PROVIDENCE, R.I. [Brown University] — “I, Robot.” “WALL-E.” “Ex Machina.” “The Matrix.” For decades, countless films and television shows have asked: What would happen if humans created robots that were so smart that they sought to replace their creators?

The question seems more eerily relevant every year, as scientists develop robots that are intelligent enough to drop fatal, precisely targeted bombs, carry on complex conversations and solve difficult math problems. Yet outside of movie theaters and scholarly communities such as Brown’s Humanity Centered Robotics Initiative, few are reckoning with the potential consequences of creating such advanced artificial intelligence.

Enter Sydney Skybetter, a senior lecturer in theatre arts and performance studies at Brown University. An experienced choreographer with an interest in artificial intelligence, Skybetter has long explored what human movement and performance can teach people about responsible robotics.

“More than ever, companies like Meta and federal agencies like the Transportation Security Administration are generating technologies that define how we exist in the world,” Skybetter said. “They say their technology can help us connect and keep us safe — on the other hand, we’ve already seen what happens when we give companies and agencies carte blanche to surveil us as they do, and we’ve seen what violence this technology can inflict both at home and abroad. I think artistic inquiry can help us understand and critique these emerging technologies — not only that, but it can also confer on these technologies a much-needed humanist core.”

In Spring 2022, alongside computer science Ph.D. candidate Eric Rosen and Class of 2021 graduate Madeline Greenberg, Skybetter sought to humanize robotics via a course called Choreorobotics 0101. The course, co-taught by Skybetter and Rosen and managed by Greenberg, drew interest from a diverse mix of students with experience in computer science, dance, theater and engineering.

On the surface, the course’s objective was straightforward: Learn how to choreograph a 30-second dance using one of two pairs of Spot robots designed by the robotics company Boston Dynamics.

Yet beneath the shiny spectacle of making the robots dance, there was a lot more to unpack. Through a mix of time spent in the dance studio, in the robotics lab and in engaging discussions, students explored the kinds of tough and important questions they may confront in their careers: What are robots for, anyway? How can they improve peoples’ lives? And how can roboticists ensure their creations aren’t used to exploit or hurt people?

As part of a quest to use artistic inquiry to better understand emerging technologies, students in Choreorobotics 0101 broke into small groups to choreograph robot dances.

Ultimately, Skybetter said, he hopes the course — a direct illustration of how scholars at Brown bring disparate perspectives to confront complex challenges — helps the next generation of engineers create emerging robotics and AI technologies that minimize harm and make a positive impact on society.

“The subject of ethics and justice in technology development is incredibly urgent — it’s on fire,” Skybetter said. “I feel it’s my job to help students understand the implications of the technology we create now and in the future, because they are the future. I can’t resolve the issues we’re exploring, but my hope is that maybe they can.” 

First, do no harm

The subjects of performance and technology aren’t as disparate as they might seem, Skybetter said. Emerging technologies such as the pointe shoe, the gas lamp and the proscenium stage shaped the trajectory of ballet, a foundational dance in Western history.

Nor, he said, are dance and violence as unrelated as many would believe. Ballet began in the royal courts of Italy and France as an artistic extension of fencing — and while today fencing is considered a safe sport, it was then a means of training military troops, whom Louis XIV famously deployed to secure French supremacy in much of Europe and to establish colonies across multiple continents.

“There is no separating the dance tradition from that military tradition,” Skybetter said. “The Western dance tradition starts with colonial violence.”

Abby Perelman (left) said she had never programmed a robot before — but with Rosen’s assistance, she felt free to try and fail.

And robotics, Skybetter said, is the perfect lodestar for conversations on dance, technology and violence, because it exists at the intersection of all three subjects. Since World War II, roboticists working with militaries across the globe have acted as precision-focused choreographers, developing sequences of motions that allow robots to gather foreign intelligence, surveil domestic and foreign civilians, and drop missiles that have the potential to kill dozens of people at a time. At the same time, robotics has also become an art form: Lady Gaga’s Super Bowl halftime performance in 2017 featured a swarm of creatively synchronized flying robots.

“Take drones, for example,” he said. “They look very cool when they fly together; they’re like synchronized swimmers. But the better these robots are at moving with precision and improvising when unpredictable things happen — in other words, the better they are at ‘dancing’ — the easier it is to deploy them to inflict violence. And that leads to one of the governing questions of this course: How do we explore and understand these technologies while, to the extent that we can, not doing harm?”

That question has loomed at the forefront of Seiji Shaw’s mind for years. Shaw, who graduated from Brown in May with a concentration in computer science, has been working with robots since childhood. At Brown, he logged countless hours in the Intelligent Robot Lab, conducting research focused on ensuring safety for both robots and their programmers.

“I’ve spent a lot of time asking, ‘How do I make a robot move this way or that way?’” Shaw said. “But what this class asks is, ‘What is that robot movement being used for?’ and ‘What does that movement convey to people?’ Thinking about the ‘why’ of robotics — thinking about it from a humanities perspective — is something I hadn’t done before, and it’s something I always felt like I had to do.”

What if we combined human instinct with logic-based problem-solving to create technology that makes the world a better place?

Seiji Shaw Class of 2022 graduate

In sessions at the dance studio, Shaw said, he learned choreography theory that helped him understand how robots receive and interpret information. In the midst of a movement exercise, Skybetter asked Shaw, who had no prior dance experience, to direct a fellow student to raise and wave his hand without using the words “raise your hand” or “wave.” Shaw realized he could direct the movement step by step, like a choreographer or a programmer: Bend at the elbow, lift the forearm, lift the shoulder, open the palm, and finally bend the wrist back and forth. His directions worked, but his partner’s hand-waving movement looked, well, robotic — stiff, unsure, less warm than a human’s wave.

“The studio classes showed me how choreographers and roboticists often ask the same questions, but they’re getting at different things,” Shaw said. “They’re both creating sequences of movements, but the choreographer’s objective is to convey emotion or critique a social issue, and the engineer’s objective is to solve a problem. It’s made me wonder, what do we lose when we solve a problem without factoring emotion in? What if we combined human instinct with logic-based problem-solving to create technology that makes the world a better place?”

Shaw may well answer that question in his lifetime: After graduation, he plans to pursue a Ph.D. in robotics and eventually become a researcher focused on making robot movement robust and safe.

Learning to ‘fail safely’

Safety was top of mind for all in Choreorobotics, including Rosen, who presided over all of the course’s robotics lab sessions. Before students interacted with the Boston Dynamics robots — hulking yellow and black devices that resemble large dogs — Rosen spent a month teaching the class basic robotics concepts and terms and reviewing Spot’s safety manual. He also provided an overview of Choreographer, the software program that brings the robots to life, which Boston Dynamics generously donated for the course.

Spending a few weeks learning the basics helped students understand how to “fail safely” as they began to choreograph their robot dances in late winter, Rosen said. Students learned how to make mistakes that they and their peers could learn from while prioritizing physical safety.

“We had a robot failure last week,” Rosen said with a sheepish smile in one March lab session. “We tried to program a bourrée with the robot’s feet too close together, and it fell over. But that’s what this class is for — understanding that robots aren’t just consistent machines that never fail. Like us, they have flaws and limitations. They bump into things they can’t see in the dark, and they get confused when they’re near reflective surfaces.”

Seiji Shaw (right) said Choreorobotics 0101 helped him confront important questions about what people use robots for today and how they might be used as a force for good in the future.

Those introductory safety sessions were key for Abby Perelman, a lifelong dancer who concentrated in cognitive neuroscience and data fluency, graduating in May.  

Outside of Choreorobotics, Perelman spent part of the spring semester conducting research on creating a wearable device that could track symptoms in Parkinson’s disease patients, allowing doctors to provide them with better care. Together, the course and her research provided extensive food for thought about how to create technology that makes a positive impact — and how to ensure that that technology isn’t used for nefarious purposes. 

Perelman had never programmed a robot before — but with Rosen’s assistance, she felt free to try and fail.

“It’s a space for people to explore and try and possibly fail and watch your robot fall over and laugh,” Perelman said. “I’ve never taken a class like this before where we’re all just coming together from our vastly different perspectives and disciplines and learning and trying together.”

Skybetter admitted the course was as much a learning opportunity for the instructors as it was for the students. The course was his first attempt at translating his research initiative, the Conference for Research on Choreographic Interfaces, into an applied learning experience involving robots. But with CRCI poised to broaden its reach as a growing initiative within the Brown Arts Institute, he believes it’s the first of many. Skybetter is already working with Nora Ayanian and Stefanie Tellex, two faculty members in the Department of Computer Science, on a 2023 or 2024 course that focuses on drone swarms — their development, their history in military operations and the performing arts, and their potential to inflict harm and do good.

Navaiya Williams, who just completed her first year at Brown and is double concentrating in computer science and theater, might enroll next year. Williams, who grew up performing and participated in robotics courses in high school, had no idea what to expect when she set foot in Choreorobotics 0101 in January — but, she said, the course was a pleasant surprise.

“When I tell people I’m double concentrating in computer science and theater, they say, ‘What are you going to do with that?’” Williams said. “And to be honest, I didn’t know what to tell them — theater and CS always seemed so different and unrelated. But this class showed me avenues I didn’t know existed before, where I can use both subjects in really creative, meaningful ways. It has me thinking about new possibilities for careers related to cybersecurity and virtual reality.”

Williams must have been listening between the lines when Skybetter, on one March day, guided students through a movement exercise inside the Ashamu Dance Studio at Brown. Lying with students on the floor, he asked them to think about their physical movements before actually moving. Consider, he directed, the muscular apparatus required to lift a hand and point a finger — the pectoral tensing, the metacarpal engagement. Be attuned, he said, to the difference between the pre-perception of the movement and the feeling of the movement itself.

“Take this weird time, in this weird studio, in this weird class, on this weird plastic floor, to try weird movements,” Skybetter said. “Try and make choices that you haven’t made before.”

Original source can be found here.