Imagine you could shape a robot yourself. Ask for the exact features you want. Decide what it looks like and what it does. Picture it clearly for a second, then hold that image in your mind.
Now here is what usually happens instead. Engineers build a robot and it's technically impressive. A hospital buys it. Then everyone else, the nurses, the patients, the aides, has to adjust their day around it. The robot comes first. The people who actually live with it get asked last, if they get asked at all.
Angelique Taylor, an assistant professor at Cornell Tech, ran a study that did it the other way around. Over 14 weeks, she and her team brought together 22 people: nurses, doctors, long-term care residents, artists, and engineers. She asked them to build healthcare robots together, from scratch. Not starting with what the technology could do. Starting with what actually frustrates people inside a hospital.
They worked in Cornell Tech's MakerLAB. First sketches. Then cardboard mockups. Then full-size, working prototypes. They tested the idea across three real places: an emergency department, a sleep disorder clinic, and a long-term care facility. Here's the part worth noticing. Hallway width, noise, patient comfort, hygiene, none of it came up in early conversations. It only showed up once someone stood next to a robot-sized shape of cardboard, in a real hallway.
Three different robots came out of it, one for each room. In the emergency department, a bear-shaped robot that delivers medical kits to a patient's room before the doctor arrives. In the sleep clinic, a calm robot with soft lighting that guides anxious patients through the night. In long-term care, a robot built less for efficiency and more for company: it shows the day's schedule and keeps residents from feeling alone.
Niti Parikh runs Cornell Tech's MakerLABs. She calls the cardboard "an instrument of thought," letting people without a technical background move from watching to deciding. She also called the lab itself a third place, outside hospital hierarchy, where "mistakes are encouraged."
That detail matters more than the robots themselves. A nurse who never gets a say in engineering decisions was, in that room, allowed to be wrong out loud. So was everyone else. The designs got better because of it, and so did people's confidence that they belonged in the conversation at all.
Which brings us back to the robot you imagined a minute ago. Ask yourself something simple. In your own workplace, did anyone ever ask you that question before the new system, the new software, the new tool showed up? Most people never get asked. The nurses and patients in this study did, cardboard and all, weeks before a single robot was finalized.
The real design question was never only what the robot could do. It was whether the person standing next to it ever got to imagine it first.
Source: Stanley, G. (2026). New Approach Designs Healthcare Robots With, Not For, the People Who Use Them. Cornell Chronicle / Cornell Tech, reporting research by Taylor, A., Parikh, N., Ju, W., Bai, Y., and Han, R., presented at ACM CHI 2026. Read the article →