Image from Blutgruppe

A revolution in robot technology and machine intelligence is upon us.  It’s an avalanche that’s going to sweep away the previous tradition of working for a living. This is not a revolution you can watch from the sidelines, it will affect every American citizen who works. Our capitalist definition of work and career is going to change in dramatic and unpredictable ways. The commercial and military advantages of artificial intelligence make it a certainty. It’s time to start thinking about the new world order and planning your future.

Computer code already writes a substantial number of news stories. Most basic sports stories, corporate reports and in some places, crime stories are automatically generated. Makes sense, most of this type of reporting is just plugging the numbers into a template. A pharmacy in California is now operated by a computer program. So far, 350,000 prescriptions filled without error. Welding jobs in auto factories were replaced with robots 40 years ago. The pace of job destruction is accelerating for both blue and white collar workers. Many supervisory and management occupations will be in the next wave. Technology has reshaped the workforce many times before, entire career categories have disappeared, from buggy whip makers to typists. So far it’s always been a long term positive factor. But is this particular cycle going to be more of the same, or is this situation different? And if this is different, how are we going to deal with it?

Let’s get some terminology out of the way first. For the purposes of this article, I’ll use these definitions.

Expert system. A device or software that can deliver solutions to problems in a narrowly defined area. A program that can diagnose medical conditions, for example. A very good expert system would have a form with boxes to check for symptoms, and a complete medical history. It makes a diagnosis by comparing symptoms against a disease database. It’s only as good as the data you feed it.

Intelligent system. For the moment, let’s keep this one simple. You go to the doctor’s office and tell the Doctor Friendly computer, version 2.0, that you don’t feel good. Doctor Friendly is not necessarily an android, but it might speak in natural language and record video. When Doctor Friendly interviews the patient, it might read the patient’s body language and decide that he’s downplaying his symptoms in order not to miss work. Or notice a slight limp or facial tic that the patient failed to mention. It would thus probe for more detail, or order additional tests.

It’s a matter of efficiency. It’s possible to feed an expert system the data from every diagnostic test imaginable, including a complete toxicology and blood chemistry screen. That would be a very expensive way to do things. An intelligent system would be able to recognize subtle symptoms or unusual behavior that would not show up on the form a diagnostic expert system uses for input. So it hones in on a short list of appropriate questions to ask and tests to run. The correct diagnosis is achieved with minimal effort and cost. Elegance as opposed to brute force.

Futurist tech prognosticators like Ray Kurzweil believe that these types of intelligent systems will be commonplace in 20 or 30 years. Seems like a big leap. The best machines we have today are fast, but dumb. Chess playing Big Blue for example. Or Jeopardy winning Watson, a much more complex system. It does a reasonable job of deciphering spoken language. But it’s still a million miles from being able to read your mood. Or understand how you feel about losing to a box full of integrated circuits.

Watson might be state of the art now, but ask yourself what a mobile phone was like 30 years ago. How many people saw the future of mobile phones ? Virtually all the predictions seriously underestimated how fast and far reaching the changes would be.

The Glass Half Full Scenario

The glass half full camp says all these newly displaced workers will go back to school, get retrained for newer, more technical jobs, and live happily ever after with higher pay.

They will cite the typing pool of the 1950s. Word processors made office work more efficient and freed women from boring and repetitive work. Very true. IT departments didn’t exist then. Neither did the Google, Facebook, or Starbucks hot spots. Word processors and spreadsheets simply made humans more productive. The theory is that new industries will rise and create millions of new jobs. It is a good point. Just because we can’t see the shape of a new economy, it doesn’t mean it can’t happen.

The Glass Half Empty Scenario

The obvious question is what happens when more than half of the workforce are unemployed? That’s the prediction of the 2013 Oxford Study, which found that 47% of all US employment was at risk of computerization within the next decade or two. When we get to the ultimate endgame, the numbers might be much higher. My own guess is that if we look ahead another decade, out to 30+ years, 80% unemployment is not out of the question.

What does life look like if 80% of the population has no income? Maybe there is some kind of underground economy where people grow veggies in their backyards or sew quilts to sell to rich people. More likely a lot of people would be manufacturing and using various illegal and dangerous drugs. Think that is ridiculous? A look at the favelas of Brazil or the barrios in the Philippines will tell you what life looks like when there are a lot of people and not enough work. Then again, there will probably be robot cops to keep order.

A New Model

Intelligent systems create an entirely new economic model. Here’s the definition of capitalism from Wikipedia.

Capitalism is an economic system in which trade, industries, and the means of production are largely or entirely privately owned and operated for profit. Central characteristics of capitalism include private property, capital accumulation, wage labor and, in many models, competitive markets. In a capitalist economy, the parties to a transaction typically determine the prices at which assets, goods, and services are exchanged. Wikipedia

With good expert systems, and later with intelligent systems, wage labor will all but disappear. Companies don’t just own the factories and machines, they own the mechanical workers.

Forget typing pools, one intelligent system would replace a whole department that creates and manages documents. Including all the managers and supervisors.

Are you thinking that someone has to make the decision about what documents should be created in the first place? That’s precisely the difference between an expert system and an intelligent system.

Of course you need someone to maintain and troubleshoot the machines when something goes wrong. Keep in mind that something going wrong will be a rare occurrence, and that person can be outsourced.

Our current economic system must morph into something very different. That doesn’t automatically mean most Americans will live in some dystopian ghetto. The Swiss for example, are considering a plan to guarantee every citizen a minimum income of $2800 per month, whether they choose to work or not. That plan also does away will all the current welfare plans, and all the bureaucracy that goes along with those programs. Imagine how that might work in the US. Food Stamps… gone. Social Security disability, also gone, unemployment insurance, does not exist. All the retraining programs, not needed any longer. All the tens of thousands of government jobs that are needed to run the various welfare and unemployment programs at both the state and federal level, also gone. It might actually be cheaper to just guarantee a reasonable income.

It’s something to think about.

I’ll post “Part 2: The shape of Things to Come” in a few days.


The Oxford Study: The Future of Employment

When robot reporters write the news – New Scientist

Ray Kurzweil and the Artificial Intelligence Singularity – The Guardian

Robots will take your job and kill the economy – FactsCoExist

The Swiss Plan to End Poverty – Bloomberg