Finding common ground to help robots master the mundane

Today we’re launching the second video in our Open Source Stories “Common Connections” series, “Common Connections: Making Robots Boring.” The series features scholars, CEOs, educators, and engineers who've never met before coming together to find the common threads in their work, and exploring the potential for future open source innovation and building unexpected connections.

“Common Connections: Making Robots Boring” brings together Leila Takayama, Ph. D., human-robot interaction researcher at the University of California, Santa Cruz, and Chris Nicholson, founder and CEO of Pathmind. We first met Takayama in part four of our film “How to Start a Robot Revolution,” when we explored how her work with the startup Willow Garage made a foundational piece of robotics software more user-friendly. While Nicholson showed up in our 2017 film, “Road to A.I.,” which featured his work creating deep learning software tools that have contributed to the growth of autonomous driving.

We’ve invited Takayama to share her thoughts on the potential of a robot uprising and why it’s important that robots stay boring.

I love boring robots. I want my robots to do their jobs without being divas. Our fancy, “automatic” coffee machine at Willow Garage was constantly demanding our help, using imperative text on its little screen. Fill beans! Empty tray! (Do it now, human!) I have enough devices to care for -- charging them, cleaning them, replacing broken parts -- without them also trying to be on centerstage all the time.

Robots hiding in plain sight

I much prefer those robots that are quietly hiding in plain sight, doing their jobs. They wash our dishes (once we load them up) and maintain a comfortable air temperature in my house (once we show them what “comfortable” means to us). Those humble devices are so boring that we don’t even bother to call them “robots.” Why can’t more robots be like that?

One possible reason is that the incentives around robotics research and development just aren’t often aligned with producing boring robots. In robotics research and development, we talk about the pressure to “demo or die.” There is some truth to that melodramatic turn of phrase. If a robotics researcher is going to “survive” in this career path, it is critical to learn how to produce engaging, memorable robot demos.

“Have you seen that Boston Dynamics video?!” That is one of the top questions I get asked whenever someone learns that I do human-robot interaction research. Clearly, the team at Boston Dynamics has mastered the art of producing engaging and memorable demo videos that show off their technical prowess, inspiring (or terrifying) people around the world.

The robot uprising? Not very likely 

That brings us to the second question I am asked most often: “Do you worry about robots taking our jobs (or taking over the world)?” My response to those questions is roughly the same every time. Of course, I’ve seen the latest Boston Dynamics video! Yes, it’s amazing. No, I’m not worried about a robot uprising because robots are not nearly as smart or capable as movies or media make them out to be

Given what people are watching on YouTube, reading in sci-fi books, and scrolling through on their media feeds, it’s not surprising that they’re worried. However, the more time you spend with real state-of-the-art robots, the more you realize how far they are from those science fiction robot villains. 

Sure, you’ve probably seen legged robots walk up and down stairs, but did you know that most of those robots need massive harnesses to hold them up for those (very common) situations in which they fall? Do you know how many times these robotics teams have to shoot that video before they finally got the one take in which the robot succeeded? 

Did you notice the person in the background with the laptop, who was not only monitoring the robot’s vitals, but was sending commands to the robot throughout the demo? Some robotics teams are more transparent than others about the many human robot wranglers that it takes to run any of those demos. I think we would all benefit from getting to peek behind the curtains so that more people could appreciate the technical prowess without being confused about how much of the performance was actually done autonomously.

These flashy demos are produced to impress potential funders for researchers, investors for startups, customers for robotic products and services, etc. I certainly empathize with the need to gather support for a vision of what robotics could be so that we can invent that future. 

More boring, please

However, I think there is also value in making robots more boring so that we might actually be willing to put up with them being around us in that future. The novelty of interacting with robots wears off pretty quickly. What we really want is technology that enhances our lives and blends into the background. 

For more information about Open Source Stories and to watch the full video, “Common Connections: Making Robots Boring,” visit Open Source Stories. Then join the conversation on social media using the hashtag #opensourcestories.