Innovation and Robotics: The Uncanny Valley of Death

The Jetsons premiered almost 50 years ago, but where’s Rosey?

The Jetsons premiered almost 50 years ago, but where’s Rosey? Science fiction has been predicting the rise of robotics in daily life for decades and, for the most part, science fact has delivered, with ever-improving artificial intelligence driven by faster processors, and continued advancements in mechanical engineering allow machines to perform increasingly complex physical functions. The White House National Robotics Initiative announced last year that “robotics technology is reaching a ‘tipping point’ and is poised for explosive growth because of improvements in core technologies such as microprocessors, sensors, and algorithms.” Still, outside of very limited and niche uses, such as assembling machinery on factory floors or vacuuming dirt from living room floors, autonomous machines have little impact on our daily lives, and the much-promised robot revolution (or apocalypse) still seems a long way off. One way to explain robots’ failure to launch is through the study of other disruptive technology. For the moment, there are several innovation trends working against them – they are stuck in the Uncanny Valley of Death.

A “funding gap” or “Valley of Death” is a firmly established concept in investment and innovation studies. Simply put, all innovation begins with basic research, which aims to increase understanding and is undertaken out of curiosity rather than a specific commercial application. While basic research doesn’t generate revenue on its own, it is typically well funded by universities looking to advance knowledge and government initiatives which recognize its vital role in advancing technology. Research only begins to turn profitable on its own during commercialization, when new discoveries become products and can be monetized. Between these stages, however, there is a point where government funding tapers off but risk remains too high and rewards too obscure for private funding to kick in. While various iterations of this progression have been proposed from different perspectives, this point in research is considered the “Valley of Death” because so many potential innovations fail here, no longer cutting edge enough for university and government support but not yet sufficiently viable for profit-maximizing corporations.

In many ways, this is where robotics is floundering. We don’t currently have the technology to build Rosey, the Jetson’s robot housekeeper, to be affordable or useful. Yet we can build highly mobile, highly functional humanoid robots and have developed fairly advanced artificial intelligence and machine learning. The closest we’ve come to a robot housekeeper, however, has been the Roomba, while sporadic attempts at humanoid household robots have been more novel than practical and prohibitively expensive. The same is true for most of the applications for robots we see in fiction. Unfortunately, the only way to develop robots that are functional and cost effective is to keep developing ones that aren’t, which government, academia, and industry are rarely willing to fund.

Another effect at work is the “Uncanny Valley”, which is hypothesized to account for the discomfort humans beings feel around realistic robots and animations. While we’re comfortable with robots that are clearly mechanical and look nothing like us, when we see artificial beings that seem almost alive, we don’t like it. Though this is typically used to refer to appearance, I believe it also applies to behavior. We don’t like it when robots perform human functions, even when they are fully capable. This can be seen in some of the poorly defined outrage at drone warfare. Drones are neither autonomous not unprecedented, raising few moral issues that technologies such as cruise missiles have not brought up, and yet have seen a massive domestic moral backlash. In part, this is because we are uncomfortable with robot planes that look and behave like manned ones, even bombing human targets. The same goes for proposed future functions. Robot doctors, for example, could allow the worlds best specialists to operate on patients from across the world, but few would currently opt to be cut open by a robot arm, no matter who was controlling it. Would you let a robot watch your kids or take care of your grandparents? One of the reasons robots have stayed on the factory floor is that this most want to keep them there.

The Department of Defense, however, still believes in the potential for robots to do better work faster and cheaper for America, and has recently launched two programs to bridge this Uncanny Valley of Death. The Defense Advanced Research Project Agency (DARPA) has announced a “grand challenge” for robot builders offering a $2 million prize for robots that can use human tools for disaster response. While a human form isn’t required, DARPA offers a humanoid platform to entrants. Are we ready for robot first responders capable of climbing ladders, navigating over rubble, and using hand tools? If DARPA’s grand challenge makes you uncomfortable,  the Naval Research Laboratory opening a  Laboratory for Autonomous Systems Research (LASR) will terrify you. Among the projects LASR hopes to pursue are autonomous robot firefighters and unmanned underwater vehicles.

Government funding is the typical method for overcoming gaps in investment caused by market forces (the Valley of Death) and biases (the Uncanny Valley) but it’s insufficient alone. For robots to take off as the likes of Isaac Asimov predicted, they will need to stand on their own or else be stuck in the same rut as alternative energy, which depends on heavy subsidies and tax breaks. Only time will tell whether robotics will overcome the Uncanny Valley of Death to become a real disruptive technology.

This post by AlexOlesker was first published at CTOvision.com.