How Robots Will Work with Us Isn’t Only a Technological Question

Many questions remain unanswered about how humans and robots will interact, including in the workplace. We just know that many more of these interactions will be taking place, as robots continue to play a greater role in our lives. Missing from many of our conversations about robots is the role of human robotic interaction (HRI) – we tend to focus on what robots can do, more than how we will work with them.

This question is critical, and its answer is ultimately dependent on questions of design. How we design our robots will shape how we work with them. In thinking about how to design our robots, a service framing is beneficial: who are a robot’s customers? The answer goes beyond those directly involved in interacting with the robot to include additional stakeholders such as families and supporters, medical staff, trained professionals such as dieticians and personal trainers, restaurants and food service providers, and even policy and lawmakers. The presence of robots will change how humans behave in ways we don’t yet fully understand, and the choices we make in designing them will help determine how.

A few years ago at Carnegie Mellon, we developed a robot, the Snackbot, along with a snack delivery service, to explore these questions. A service design framing helped us to develop the robot holistically, rather than merely seeking to advance autonomous technology. Stakeholders included customers, others in the workplace, the robot developers, designers, and researchers, the robot’s assistants, and the people who obtain and load the snacks on the robots. The context of use and the norms of the workplace also needed to be considered. A product-service system was developed to track information about behavior and preferences over time. Knowledge about how the technology influenced human behavior and how human behavior influenced the technology was used to tune the robot’s design.

Our findings, published in 2012, only confirmed how greatly design decisions weigh in shaping our behavior when dealing with robots. First, we found that that subjects anthroporphized the robot, saying things like “Snackbot doesn’t have feelings, but I wouldn’t want to just take the snack and shut the door in its face.” Second, we found that the presence of the Snackbot caused both positive and negative ripple effects in the workplace, leading to new and different interactions between coworkers as colleagues observed each others’ interactions with the robot.

Finally, we experimented by personalizing Snackbot, so that it would speak to participants drawing on knowledge of their previous interactions. In many cases, this seems to have deepened participants’ interests in interacting with the robot over time. We saw signals of trust and rapport, for example when people dressed the robot up in hats and beads, and when a customer brought Snackbot a snack — a battery — after the robot died one day in front of her office. In another case, the robot’s commenting on subjects’ snacking history led to discomfort from subjects who preferred not to be reminded of the junk food they ordered the previous day. Clearly, the choice of whether or not to personalize the robot’s service affected how humans interacted with it.

As we consider the role that robots will play in our offices and in our lives, we must remember that their capabilities are not simply defined by the cutting edge of technology. They are also the result of the design choices that we make.

This entry was posted in Leadership, Technology. Bookmark the permalink.

Comments are closed.