Tuesday March 28, 2017 0 comments
By Thomas Frey
The DaVinci Institute
Over the past few weeks I’ve come across a growing number of tirades warning us of the dangers of robots. If you were concerned about A.I. in the past, once it’s tied to a robot, the warnings seem to escalate to imminent peril. Each of these writers seems to reach the same conclusion, that super smart robots will soon be taking all of our jobs, and humanity is doomed.
The biggest misconception about AI is that if we create intelligent systems, those intelligent systems will want to overthrow their human managers and take over the world.
We’ve seen this many times in the movies — evil robots taking over the world, where technology is the bad guy, and only Jeff Goldblum can save us.
A recent survey by SelectHub showed that 41% of Americans fear getting replaced by AI, automation, and digitization. This fear plays out in different ways depending on age, gender, and social status.
A full 50% of Gen Xers have concluded they will need to get a job in a different industry if their current position gets eliminated through AI, automation, or digitization.
The victimization mindset continues to grow as people worry about becoming irrelevant.
As a result, we now have a burgeoning tech-wary community, fueled with paranoia, ready to halt or at least limit progress before it even leaves the starting blocks.
Recently Bill Gates suggested that robots should be taxed to offset the cost of humans losing their jobs:
“Right now, the human worker who does, say, $50,000 worth of work in a factory, that income is taxed and you get income tax, social security tax, all those things. If a robot comes in to do the same thing, you’d think that we’d tax the robot at a similar level.”
While on the surface it would appear to be logical response to the almost daily announcements about jobs disappearing, and an easy campaign slogan for politicians, this line of thinking is problematic on many different levels.
Taxing Robots 101
Naturally, when it comes to taxing robots, everything centers around how we define them. The dictionary definition is rather broad, describing a robot as “a machine capable of carrying out a complex series of actions automatically, especially one programmable by a computer.”
If we use this definition, then robots will include everything from cars, to printers, elevators, clocks, tractors, forklifts, guns, lawnmowers, chainsaws, drones, and much more.
If we think of robots as job destroyers, we need to understand that software is far better at it than machines. Case in point, sharing economy companies like Uber and AirBNB employ sophisticated apps that have decimated both worker and management jobs alike. Even simple spreadsheet software can reduce the time an accountant spends doing company books by over 90%.
Also, there’s a very fine line between a machine that is laborsaving and one that is job-killing. Nearly every patent filing includes at least one laborsaving claim in its extended description and this has been going on for hundreds of years.
Limits to Automation
If we think about automation 1,000 years from now, what are hard limits? What things are possible and what’s not?
So far we’ve been limited by technology and something called the Polanyi paradox, named after Karl Polanyi, an Austro-Hungarian economist who in 1966 concluded, “We know more than we can tell.”
His paradox refers to the difficulty in automating an activity that we only understand implicitly like painting a picture, writing a persuasive argument, or dancing. All of these are tasks that even people who are highly skilled in them are not fully able to describe.
We can’t automate what we can’t understand.
While we have seen some cases where machine learning appears to be capable of understanding our implicit capabilities, for now, professions that require resourcefulness, flexibility, and creativity still appear to be impervious to obsolescence.
14 Fallacies of the Coming Robot Apocalypse
Here are some of the biggest robot myths floating around today:
1.) Robots are destroying jobs. Wrong. Robots don’t eliminate jobs, only tasks.
The first misconception is that automation destroys jobs, which is not true. It does kill parts of jobs and eliminates the needs for certain skills, but entire jobs are far more complex than that.
2.) Automation has already destroyed many jobs. Wrong. Automation has only completely eliminated one job in the past 67 years.
According to a recent report by Harvard economist James Bessen, automation has only caused one job to go totally extinct over the past 67 years – elevator operators.
Of 270 occupations listed in the 1950 US Census, only elevator operator no longer exists due to automation.
Another 32 jobs were done in by a loss of demand, and five became technologically obsolete.
3.) Automation will soon eliminate my job. Wrong. Automation is forcing companies to redefine jobs.
In the short to medium term, the main effect of automation will not necessarily be eliminating jobs, but redefining them. As the skills and tasks required in the economy change, our response should not be alarmism or protectionism, but a strategic investment in education
ATM machines did replace many of the tasks that bank tellers performed, but not all of them. As a result, ATMs enabled tellers to be more efficient doing other things.
Similarly, meter readers do far more than read meters. Retail clerks do more than operate cash registers. And travel agents, teachers, truck drivers, tollbooth operators, and even parking lot attendants are all more than single task jobs.
While automation can drastically reduce the number of people needed to perform a specific job, the job itself is rarely eliminated.
4.) Automation is reducing our opportunity for finding a job. Wrong. Automation often increases the number of jobs.
The textile industry is a great example of this phenomenon. Despite the fact that 98% of the functions of making materials have now been automated, the number of weaving jobs has increased since the 19th century. As automation drove the price of cloth down, the lower prices increased demand, and eventually caused more job growth.
In a similar fashion, when ATMs rolled out in the 1970s, their numbers grew rapidly from 100,000 to 400,000 between 1995 and 2010. Since operating an ATM is cheaper than paying a teller’s salary, we started seeing more ATMs than tellers, and the overall cost of operating a bank branch came down. Since it was cheaper to operate a bank branch, more of them opened, and the number of bank branches increasing by 40% between 1988 and 2004.
As a result, the number of tellers actually increased. Rather than putting tellers out of work, the number of tellers continued to increase between 1980 and 2010.
5.) With automation, there will be nothing left for humans to do. Wrong. Non-automated tasks will become more valuable.
Automation is more likely to take over boring and repetitive tasks, allowing skilled workers more time to do the things that require talent.
In an emergency room setting, if diagnosis can be automated, doctors can focus on special one-off cases, increasing the overall number of patients treated.
Likewise, as automation helps mortgage-loan officers do the routine paperwork involved in processing loan applications, each person can both manage and process more loans, and their overall value to the organization increases.
6.) There will soon be a robot knocking on my door to take my job. Wrong. Robots don’t eliminate jobs, people do.
The number of workers needed is always a management decision. It’s easy to start blaming robots for the decisions made by their owners. But robots need owners.
Yes, it may be possible to construct autonomous AI robots in the future that can operate independently without humans anywhere in the picture, but that will be in a distant future, and in all likelihood, many things will go wrong.
From my vantage point, it’s very difficult to imagine a robot that is capable of taking initiative, and continually develops and redevelops its own role, purpose, and mission independent of any human agenda.
Keep in mind that with the cars we’re currently driving, it’s taken 120 years of reimagining them to get to the vehicles we have today. Even though things are speeding up and we are going through exponential growth curves in product development cycles, the kinds of robots we’re imagining are exponentially more complicated than any manmade device so far.
7.) A conscious robot means that they will work and act exactly like humans. Wrong. A.I. will not add consciousness to robots.
At this point we don’t even know what consciousness is let alone integrate it into artificial intelligence.
8.) An intelligent robot will have all the same feelings as humans. Wrong. A.I. will not enable robots to have human-like emotions.
Emotions are the effect of low-level/instinctive drives and the anticipations of rewards. They are the mechanisms we use to place value on the objects around us. Yes, we can replicate emotions on a certain level, but artificial love will still lack many of the quirky trace characteristics that make us human.
9.) Intelligent robots will want to overthrow the human race. Wrong. Robots do not come with a “gene” that causes them to desire power.
No they don’t. This seems to be a reoccurring theme in countless Hollywood movies, but few are asking the most relevant question of “then what?” If they somehow manage to conquer humans and only machines remain, then what?
10.) Super smart robots will be thinking machines. Wrong. Robots are incapable of human-like thinking.
Naturally this depends on your definition of thinking, but the human brain is extraordinarily complex, with around 100 billion neurons and 1,000 trillion synaptic interconnections. The brain is not digital. Rather, mental capabilities are dependent upon electrochemical signals with inter-related timing and analogue mechanisms, the sort of molecular and biological machinery that we are only just now starting to understand. Simulated thinking is still a long ways away in robots and will be a far cry from the way humans think.
11.) Robots will soon be competing for your job. Wrong. Robots don’t compete.
Robots don’t come with a built-in desire to compete. They only do what they’re told. Yes, in many situations they can perform better than you, doing the same task only faster and more efficiently, but they aren’t the ones making the decision about whether you should stay on as an employee.
12.) Robots with A.I. will soon be able to solve all of our problems. Wrong. Robots create more problems than they solve.
Yes, some of the emerging A.I. systems will be able to solve some of our problems some of the time, but they will also create new ones. Every machine wears out. Every computer system eventually dies. While the mean time between failures will undoubtedly get longer, they will all inevitably fail.
13.) We will always be able to tell the difference between humans and future robots by peeling back their skin. Wrong. Future robots will be grown just like humans.
I only included this because it’s a common theme in today’s movies. Future robots will likely be cloned or “grown” into living breathing fleshbots, so cutting off part of their arm will just reveal a very humanlike severed arm with blood gushing everywhere.
14.) Robots are forever. Wrong. The second law of thermo dynamics states that everything is slowly falling apart, even robots.
All robots, computer systems, AI software programs, and their synthetic hands and legs will eventually fail. This means that someone will need to be there to make repairs, pick up the pieces, and do the scheduled maintenance.
Yes machines can fix machines just like human fix humans, but at some point, all systems eventually fail, just as humans do.
Job paranoia is running rampant. We have seen what machines can do and it’s making us all very nervous.
However, it’s important to understand that if a profession is completely automated, yes, jobs will ultimately be eliminated. But if the process is only partial, eliminating only tasks rather than entire jobs, employment for those jobs may in fact increase because of the efficiency gains and possible effects on demand.
So far we’ve managed to eliminate many tasks, but not many jobs in their entirety. Fewer than 5% of the jobs in the U.S. today can be completely automated using current technology.
Yes, we will have to learn many new skills to stay relevant and competitive in the future, and future robots will be able to do things we never dreamed possible. But for now, there are many areas where tacit skills create safe ground for future employment.
One of my future columns will be dedicated to tasks that are not easy to automate, because this is where most people will want to hone their skills.