Kinetra Robotics occupies two floors of a former printing works in Seongsu, the Seoul neighbourhood where old factories have become cafés and design studios. On the upper floor, engineers test robotic arms that sort medicines in hospital pharmacies and fold towels in hotel laundries. On the lower floor, three professional dancers spend their days watching those arms move, and telling the engineers what is wrong with them. I asked the founder, Seo-yeon Baek, to explain the arrangement.
The problem, she said, is that most industrial robots were designed for cages. They move along the most efficient path between two points, which is often abrupt, unpredictable and faintly menacing to anyone standing nearby. That was fine on a car assembly line, but her machines work alongside nurses, pharmacists and hotel staff. In early trials, people simply refused to stand near them. The robots were safe, she said, but they did not look safe, and that turned out to matter just as much.
Why dancers, rather than psychologists or interaction designers? Baek, who studied mechanical engineering before spending two years in a contemporary dance company, laughed. Dancers, she said, spend their lives thinking about how movement communicates intention. A good dancer can signal where they are about to go before they go there, and an audience relaxes because it can anticipate them. Her choreographers design what she calls legible motion, movements that announce themselves slightly in advance.
Designing for the nervous bystander
I asked how that works in practice. The arms now pause fractionally before changing direction, she explained, and they slow as they approach a person rather than stopping abruptly. They follow curved rather than straight paths, which humans read as more deliberate. None of these changes improve efficiency on paper. Some make the robots measurably slower. But in hospital trials, the time staff spent waiting for machines fell sharply, because people stopped hovering nervously and simply got on with their own work.
An engineer asks whether the robot is safe. A choreographer asks whether the robot feels safe. You need both answers, and for years we only asked one.
Managing engineers and dancers in the same company has not always been harmonious, Baek admitted. In the early days the engineers regarded the choreographers as decorative, and the choreographers found the engineers literal-minded. What changed things was a rule that every new motion design must be reviewed with both groups in the room, watching the same footage. Disagreements are now settled by trials with real hospital staff, and the results are shared with everyone, including the losing side.
Kinetra now employs sixty people and has installed machines in hospitals across Korea and in two hotel groups in Japan. Baek is wary of describing her approach as a trend. Most robotics companies, she said, still treat human comfort as a problem for the marketing department. She would like that to change, but she is not holding her breath, and she regards the gap as a competitive advantage for as long as it lasts.
Before I left, Baek took me downstairs to watch a rehearsal. One of the dancers was walking slowly towards a robotic arm while it reached for a box, and the two of them seemed to negotiate the space between them without a word. It was oddly moving to watch. Baek smiled at my reaction. That, she said, is the whole product. Everything else is just motors and very good software.
