Skip to main content

Robotics Expert Lionel Robert on the Future of Human-Robot Interactions and What’s Next

Lionel Robert. Credit: Jeffrey M. Smith - University of Michigan School of Information.

For researcher and professor Lionel Robert, the difference between science fiction and reality is a matter of time — and “time is catching up with us.”

Countless books and movies have explored various fictional scenarios featuring robots, some of them downright terrifying. Now those fantastical artificial humans are here; the actual age of the robots is in its infancy.

But there’s still a lot we don’t know about them.

Robert, associate professor of Information at the University of Michigan, School of Information (UMSI), and an expert on human-robot interaction, is one of the people working to fill in the blanks. He studies how people interact with technology, from robots to self-driving cars to online platforms. For Robert, it’s critical to consider our needs and the broader implications of artificial intelligence (AI) now, while capabilities are still being developed.

While Robert, who studies collaborations between humans and machines, is optimistic about our future with artificial intelligence, he says, that's “not because I believe in technology but because I believe in people. We have a voice, and I think as long as that's there, we're going to move in a positive direction.”

The word “robot” was introduced in the 1920 science fiction play “RUR, or Rossum’s Universal Robots,” by the Czech writer Karel Čapek, as a riff on “robota,” a Slavic word for servitude.

In the past, industrial robots were confined to cages for the safety of the humans in the room, but robots that are safe and relatable enough to actually collaborate with humans — cobots — will be important to the future of the field, Robert says.

Robert did a study a couple of years ago that showed that the more human-like a robot is, the more willing people are to work with it — and even perhaps to prefer the robot to a human co-worker. Similarities, both visible ones like gender, age and race, as well as deeper traits like knowledge and attitude, can warm people up to robots, too. In human-robot interactions, “birds of feather flock together,” Robert says.

But making robots appear to be more like people can affect our emotional response to them, as well as our expectations of what they can and should do, he says. One example Robert cites is of soldiers who become so attached to bomb disposal robots that they resist deploying them, even though the robots were designed to spare humans from performing that dangerous job.

People often fear robots because they replace human workers, Robert says. Based on his research and analysis, he says robots will continue to take jobs, but notes that they can easily do many of the irksome, filthy, tedious and dangerous jobs humans have done up until now.

And, “they can make a lot of things cheaper, like CT scans,” Robert says. But robots don’t do well with tasks that require creativity, he says.

In general, trust is crucial between humans and AI, Robert says. As director of the Michigan Autonomous Vehicle Research Intergroup Collaboration (MAVRIC), he has studied how people come to trust autonomous cars. As autonomous cars are being hailed as a way to make roads safer, how we view these cars, and how likely we are to use them, is increasingly important for a researcher like Robert.

“Trust is based on expectations,” Robert says. Driving is all about independence and control, he says. “We’re asking people to hand that over to a vehicle. There’s not only an issue of how reliable the technology is; there is some psychology” involved in people learning to let a car do the driving, he says.

And the technology is still progressing. A big challenge for autonomous cars is the unpredictable behavior of pedestrians, bicyclists and other random humans — including handicapped people, whom, Robert says, designers don't seem to take into consideration.

Robert says the consensus now is to design autonomous cars for the most common scenarios and address so-called “special” cases later.

“But we know that doesn't work very well. I've been advocating that people consider disabled people now, not later,” he says. “We can get 98 percent of machine learning right, but the other 2% will kill you, literally.”

On the other hand, Robert’s work with online platforms offers valuable insight for those who are socially distancing now, or stuck at home, using Zoom and Teams to chat with colleagues or students. One study he did, of diverse groups working remotely, showed that texting worked best if group members were racially diverse, while video meetings got the best outcomes with gender-diverse groups. The team concluded that race tends to recede in online collaborations, but gender differences can be exacerbated if men are aggressive or “tone deaf.” Seeing women in the meeting can help men rein in those impulses.

Robert’s expertise with online collaboration has made him a valuable member of a small group of scientist AAAS “Superheroes” that work with AAAS — in his case, advising on how to engage members interactively on the AAAS Member Community.

He says the project could involve getting members to connect online, through presentations or classes, for example. “We want to encourage more participation,” he says.

Date
Blog Name

Author

Delia O'Hara