These may be relatively good economic times, with official unemployment about as low as it can get during a permanent jobs crisis, but that hasn’t stopped advances in the implementation and theory of robotics. What’s happened over the past five months?
In Financial Times on May 3rd, the title of Sam Fleming’s article, “Why robots are coming for US service jobs,” means it could contain only two words: They’re cheaper. Fleming addressed more than that, though, with a good rundown of positions susceptible to replacement by automatons, and correctly showed that the jobs now most at risk were white-collar and caregiving ones. He also cited a McKinsey Global Institute study claiming that 40% of American workers had occupations in which half their hours went to tasks that could already, with current technology, be automated. Another piece in the same publication that day, “Rise of the robots is sparking an investment boom,” showed that while venture capital investments there had to double in 2015 just to reach a still-puny $587 million, the entire market projects to reach $135 billion in three years. That’s not as massive as it could be either, and both numbers are and will be exceeded by those for driverless vehicles.
Two days later, Financial Times continued its series by asking us to “meet the cobots,” automata designed to lighten loads for existing workers instead of replacing them. Author Peggy Hollinger seemed to imply that, as a result, robots now won’t cost jobs. This idea is nothing new; when I visited a Florida postal sorting center over 15 years ago, a large yellow one moved heavy packages around alongside dozens of human workers, and precipitated anything but hostile reactions, as our tour guide told us that “everyone likes Big Bird.” When robots only assist, they serve as tools similar to computers, copiers, or even brooms, so there’s nothing special here on that count either.
On May 14th, Phil Torres weighed in in Salon with “Fear our new robot overlords: This is why you need to take artificial intelligence seriously.” He started with the Terminator movie series, which introduced me, for one, to the idea of autonomous goal-seeking devices automatically having potential problems, and showed how artificial general intelligence, or AGI, could, as happened in the first Terminator feature, destroy many or all humans if we fail at “making sure their values,” not just their objectives, “align with ours.” This material is well worth reading, especially for those not yet familiar with it.
Another Financial Times article, May 16th’s “Legal firms unleash office automatons,” reported something old as if it were new. Automated legal searches were contemporary enough for me to cite them five years ago in Work’s New Age, and while they are now getting better and more common, their changes seem only incremental. The idea of “Uberisation,” or more work being done by lower-paid workers, is nothing fresh either, with at least a strong foothold in law long before that label would have been understood.
In Harvard Business Review, Vasant Dhar’s ambitious May 17th “When to Trust Robots with Decisions, and When Not To,” presented a grid of products and needs positioned by predictability (high = fighter drones and cataract surgery; low = stock trading and effectiveness of online advertising) and cost per mistake (high = driverless cars and diabetes prediction; low = spam filtering and early education support). Since the best areas for automated solutions are clearly those with high predictability and low cost per mistake, progress will move from those toward the other corner. Dhar’s Decision Automation Map, showing this and more, is a fine tool, and he succeeded admirably at showing where we might best concentrate upcoming robotic and computer efforts.
On May 25th, Fox News published “Pizza Hut rolling out robot servers in Japan.” They cost only $1,600 apiece, and offer “a more efficient dining experience” as well as paying for themselves remarkably quickly. Such automata are also under consideration by Carls Jr. and McDonalds, to name only two fast-food chains, and will certainly spread widely, if they have as little as mediocre customer acceptance, as minimum wages increase.
The Economist, in June 4th’s “I’m afraid I can’t do that,” cited a Centre for European Economic Research working paper claiming that since relatively few entire jobs can be fully automated, we have “reasons to be less afraid about the march of the machines.” The major flaw in this thinking is that great cost savings motivate employers to rearrange positions by concentrating hard-to-mechanize tasks in those jobs to be still held by humans. It also makes the mistake of taking the past, where people in manufacturing positions easily found service jobs, as a proxy for the present, in which we know of no type of paid work capable of replacing them in turn.
Trevor Moss reported in the June 21st Wall Street Journal “Robots on Track to Bump Humans from Call-Center Jobs.” Why not? And it’s been happening for years – you see it when you call in with a problem and a robotic voice asks you several technical questions. A decade or more ago, most companies moved what was called Tier 1 support, or preliminary and easy-to-fix problem-solving, to cheaper-labor countries, and now those responsibilities are increasingly being covered by machines. There is no case that more and more sophisticated issues will not be handled, as time goes on, in the same way.
“Industrial robot sales hit record,” Financial Times pointed out June 22nd. Though worldwide sales reached only 248,000 last year, it is noteworthy that China, once known as a cheap-labor source, got a quarter of them. The automotive, electrical, and electronics industries are the largest consumers, and we can safely bet that number will increase tenfold within as many years. Robotic restaurant servers costing only the equivalent of $1,200 apiece are also arriving in that country, with high acceptance offsetting lower labor savings, as documented in The Wall Street Journal’s July 24th “In China, a Robot’s Place Is in the Kitchen.”
Finally, Xerox released a list of 19 current and near-future robotic innovations. They are robotic pharmacist, Japan’s robot hotel, digital nursing, robotic process automation, virtual customer service agents, self-flying planes, self-driving cars, driverless trains, digital barista, automated passport control, automatic translation, automatic report writing, legal work, Amazon’s “robot army,” robot security guard, the automated college professor, home automation, the robot bartender, and robot-assisted surgery. You can read more about them at https://www.xerox.com/en-us/insights/robotic-innovations?CMP=BAC_Repo2016&SECTN=IN_&SITE=TheAltantic_&SIZE=1x1.
After all this reporting of advancement, The Washington Post’s Robert J. Samuelson again told us, on August 17th, that “our robot panic is overblown.” He again fell into traps by saying “lost jobs and destroyed industries give way, over time, to new industries and jobs” (unless they don’t), and “if robots cut costs, the savings have to go somewhere” (into the massive, stagnant pools of money held by the largest corporations and wealthiest individuals). He partially redeemed himself by stating that “government’s main role is to maintain the conditions that make hiring profitable,” which, though incomplete, should be a worthy goal for both political parties, but in general, Samuelson, who I am certain does not personally invest only in stocks which have gone up in the past, should know better.
That’s all for now. There will be more, in this area contending only with self-driving cars for the most press and the greatest effect on American employment. I will continue to keep you up to date.