A great deal of communication is non-verbal. After all, a loving embrace or a cold stare can often worth a thousand words. This is part of why social robots still leave a lot to be desired, they’re still mostly limited to communicating through screens or with janky artificial voices that make conversation feel unnatural. But earlier this year, a team of roboticists at Cornell University took a huge step toward changing that.
To help robots learn to communicate via touch and feel, mechanical engineering student Yuhan Hu, along with her colleagues, retrofitted a cute red robot with shapeshifting skin. When it’s programmed to signal happiness, its pale white synthetic skin puffs up into small balloons, and when it’s supposed to be angry rounded spikes vigorously pulsate.
Hu told Inverse that letting humans feel its feelings not only gives bots another way to communicate, but it could vastly improve how much robot interaction resonates with humans.
This is #9 on Inverse’s list of the 20 Ways A.I. Became More Human in 2018.
“We think the interesting thing of the skin expression is that it operates on two channels at once: They can be perceived visually and also haptically,” says Hu. “This offers new kinds of interactions between robot and humans, and may cause more psychological impact and perhaps generate subconscious or unconscious interactions.”
One study published in the journal PLoS One asserts that people can recall interactions involving sight and touch more readily than interactions they simply hear. Yet more research published in PLoS One found that touching fills in the informational gaps left by seeing and hearing. This is especially true for children in their early stages of development when they lack the ability to speak, according to a study in the journal Pediatric Child Health.
The robot can express happiness, sleepiness, anger, and sadness through different shaped goosebumps and its face on a screen. Its skin is made out of elastomer — a very stretchy synthetic polymer — with tiny chambers that are pumped full of air depending on what emotion it wants to express.
These so-called “Texture Units” vary between goosebumps or spikes. Hu and her team are working on imbuing the adorable bot with more feels. It could one day be a useful tool to teach toddlers about emotional responses in the future.