Step aside, Furby. This experimental robot’s language skills are way cooler and way more useful. In a few years, it could even help deaf and hearing children learn their first language.
The device, called the Robot Avatar Thermal-Enhanced prototype (or, more commonly, RAVE) could someday hang above an infant’s crib, detect when they’re ready to learn, and then begin to communicate with them. It’s based off a 3D-printed robot, an algorithm based on neural and thermal scans, and a computer screen with an avatar that uses American Sign Language at a certain, optimal pace. It’s the product of a three-year, $1.1 million grant from the NSF and nearly $1 million of additional funding from the W.M. Keck Foundation and builds on research that Petitto has done for decades.
Preliminary results indicate that the robot is able to keep children’s attention for up to six minutes at a time. Infants who are only six to eight months old began to gesture in a rhythm linked with sign language after only a few minutes during this early work. Getting the hang of that rhythm might indicate that a child is learning some important foundations of any language, like picking out the sounds or other basic units of a language.
“This is work in progress,” emphasized Laura Ann Petitto, the lead researcher and an educational neuroscientist at Gallaudet University. She will present the preliminary results of her work with RAVE to a National Science Foundation committee on October 24. “It’s evolving and will continue to evolve for another year,” she said.
RAVE could be especially useful for children who are born deaf. Hearing parents of deaf or hard-of-hearing children face a choice: learn American Sign Language—quickly—or risk that their child may struggle to pick up language later in life.
Many linguists accept that there is a critical period in an infant and young child’s life that they’re particularly sensitive to language. Without hearing or seeing some language when they are very young, children wouldn’t be able to learn.
About two to three of every 1,000 children in the United States is born with some kind of hearing loss, though some national surveys have reported higher rates. Almost all of these children are born to parents who can hear.
“There’s no question that this work will have a direct impact on young, deaf infants and their academic trajectory through life,” Petitto said. Gallaudet, located in Washington, D.C., is the only university in the world designed for students with hearing loss.
However, she added, it could also be useful to addressing disparities that contribute to other kinds of inequality, even among hearing children. Some research has pointed to a “word gap” between children from wealthier families and those from poor ones that develops before kindergarten. (However, Petitto noted, some of the specific research showing a 30-million word gap had serious methodological issues, including small sample sizes.)
The fact that this tool could get and keep an infant’s attention in a meaningful way is remarkable on its own. But the robot seems to have overcome a major issue that has plagued other language acquisition tools. The kind of human interaction children need to learn language isn’t easily replicated. Studies on Baby Einstein videotapes, for example, didn’t find any positive effects on a child’s vocabulary.
Specifically, Petitto said, babies need interaction that actually makes sense. “It’s not good enough that there’s just the social interaction. When communicative actions happen, they have to be relevant and contingent and meaningful based on what the child just said,” she explained.
Still, don’t go looking for RAVE next to Furbies at the store just yet. It’s still a prototype, and the research is still preliminary. It also won’t replace all human interaction; even with RAVE, children will still need to communicate with humans to learn language. Though Petitto said the data is promising, it’s likely still far, far too early to tell if RAVE has any long-term effects on a child’s language abilities.
Late next year, Petitto said she and her team will probably start thinking about production. The team is planning another round of tests with a tweaked robot this month. “As scientists, we’re not going to sit back and read a book now,” she said. “We noticed a couple of things that we wanted to fix and we’re fixing them."
More from Newsweek