I've had a really good laugh lately at the people running around tripping robots.  I wonder if that's an actual job description (Robot Tripper) or if Google and others simply look for volunteers who are willing to trip the robots. Aren't the people doing this afraid that one day the robots will seek revenge?

But here's where innovation turns down a potentially blind alley.  Why do we need bipedal robots, and is being upright and bipedal a evolutionary feature that may have had an advantage in the past, but no longer?  In other words, if the robot gets tripped, so what?  We humans evolved to operate only in an upright position, but that doesn't mean that robots need to.  If a bipedal robot gets tripped, why doesn't it simply sprout wheels and go on its way?

In all seriousness, what we view now as advantages and features may eventually be bugs.  There are few activities that demand that a robot have bipedal capability or propulsion, when other forms of movement are equally viable and less subject to disruption.  We humans probably evolved in an upright position from a former crouched position because seeing over the grasslands meant earlier warnings about predators.  Since we no longer live in sub-saharan grasslands and constantly scan for predators, we are probably more likely to evolve in the future into beings permanently bent at the waist from sitting so much and permanently bent at the neck from scanning our devices.  Perhaps we should structure our future robots in the shape of a question mark rather than upright.

This is the problem with a lot of innovation thinking - anticipating a future that's based on the observable past.  This approach often fails to incorporate the amount of change, and the direction of change in the environment.  It's what militaries call "fighting the last war", building trenches when the enemy is building highly maneuverable tanks.  In the short run we may need bipedal, upright, humanoid robots simply because the physical environment we live in and work in is designed for humans to operate in, and therefore robots shaped like humans may be more efficient.  But I doubt it will take long before the environments change to adapt to robots and humans will adjust.  Already robots are delivering food and medicines in hospitals, and there are fully automated "dark" warehouses where no humans are employed, only robots.  McDonalds is exploring new restaurant designs that could be fully automated, with few or no human employees.  What's interesting about that design is that the "back office" - kitchen, freezers, etc - could be radically redesigned to save space, since humans will rarely or never go there, while the dining area would retain its human-centric design.

All of this to say that as we innovate, we need to consider two important issues.  First, innovation, especially disruptive innovation, rarely takes on the shape, contour or business model of what already exists.  Just ask BlockBuster or Tower Records.  Second, we have to consider the secondary and tertiary effects of innovation.  Do the robots conform to human centric designs or do buildings and factories become tailored to the robots that work there?  Most likely the latter, since robots don't need light, or heat, or safety equipment, or gyms, or cafeterias.  Too often we "innovate" only the most obvious piece of the equation - the evident product or service - and neglect how much change is occuring.  This evident innovation simply reinforces past ideas and conventions, while the rest of the world is leaving them behind.  Too often we innovate only a component of the total solution - developing a bipedal robot, when we actually should be rethinking the environment where a robot, regardless of its locomotion, is optimally deployed.


I'd love to end this post with a reference to playing checkers or chess, introducing the more complex dimensions of innovation, but that reference won't do.  It's too constrained.  Good innovators need to be thinking about the multi-variant possibilities of future scenarios rather than innovating based on past conventions and needs.  Further, once a potential future option becomes a reality, we must also understand how to operate in that new future.  Why build bipedal robots in an upright position when that's not optimal or valuable, but simply mimicking human construction?