Manufacturing was offshored to countries with lower labour costs, but not, initially, because of robots. Indeed robots have been developed as a specific onshoring strategy to bring manufacturing back to the US. Australian Rodney Brooks, the inventor of the Roomba vacuum cleaning robot, founded Rethink Robotics with the aim of making American manufacturing more competitive. Rethink produces the Baxter robot, a cheap, flexible, human-safe collaborative robot.
|Baxter is made in the US to convince companies that they can be competitive via onshoring|
Ironically, Bill Gates has recently called for robots who replace human workers to be taxed at the same rate that human workers are taxed. This seems fair, but how exactly will we know when robots have taken our jobs? While the US installed about 135,000 new industrial robots between 2010 and 2015, the number of employees in the automotive sector increased by 230,000 during the same period (IFR International Federation of Robotics) and there is a complex relationship between automation, employment growth and productivity (A3 White Paper).
Indeed it is automation rather than robots per se that are the biggest threat to human jobs, it's just that robots are the obvious, physical manifestation of artificial intelligence.
|Source: "Uses of AI and Machine Learning", Gartner (October 2016)|
"Taking the robot out of the human," is how Leslie Wilcocks, London School of Economics, describes robotic process automation. Repetitive tasks are not a strong point for humans so why not let robots do that work? Unlike the job losses of the past, concentrated amongst blue collar workers, labour disruption caused by AI and machine learning will impact the invisible jobs of white collar workers and some commentators are predicting a future without work.
So how does this tie in to the US election result? When offshoring began in the US in the 1980s it mainly impacted blue collar workers. An interesting analysis, by Joan Williams (UC Hastings), of the US working class responsible for the political success of Donald Trump, suggests that it is the removal of jobs in certain regions that causes social upheaval. She observes that many of the people displaced by offshoring were never able to gain employment again, despite being willing to work. Many US working class families have admirably deep, strong ties to their local community, hence moving away from those networks to find employment is not an option.
In contrast, white collar or "professional" workers tend to have more broad but shallow networks, making them more mobile and resilient to job loss. Arguably it is easier for white collar workers to move and retrain for different jobs if displaced by "robots". Does this mean we shouldn't be worried about robots taking our jobs?
Although the effect of automation on the workplace may be overstated (see Jeff Borland's sensible study "Are our jobs being taken by robots") the discontent in the UK and US over job security and wealth distribution means that we should be worried, we should be very worried indeed.
In a recent opinion piece, the New York Times argued that robots aren't killing the American Dream - public policy is. We are living in an era where earnings inequality continues to grow. Over the past 40 years in Australia, wages have risen by 59% for the top 10th of the population and by only 15% for the bottom 10th. According to labour economist John Mangan, "the pay gap is now so vast that while people may be sharing the same geographical area, they may as well be living in a different society".
I'm an optimist, I believe that we will all benefit from the many amazing technologies currently being developed in robotics, computer vision, and AI, but these technologies will be disruptive. How will we deal with the disruption? Well, I also believe that humans are endlessly inventive and can adapt to any situation, but only if we maintain social cohesion. For us to enjoy the promise of Ray Kurzweil's Singularity we must all be able to share in the benefits wrought by new technologies. Unfortunately this is not a problem that we can program robots to solve for us.