Will artificial intelligence take your job?

How AI will change your career over the next 20 years.

Kirsten Horton
May 10, 2018 · 8 min read

As a Master’s student researching artificial intelligence policy, I’ve read a lot of headlines about AI stealing jobs. People are really worried that automation will replace the majority of workers.

This concern is nothing new. We’ve been obsessed with technology taking our jobs for 500 years.

Stocking Frame Knitting Machine

In 1589, inventor William Lee traveled to London to apply for a patent. He hoped his stocking frame knitting machine would make hand knitting obsolete. Queen Elizabeth I was not impressed. She replied:

Thou aimest high, Master Lee. Consider thou what the invention could do to my poor subjects. It would assuredly bring to them ruin by depriving them of employment, thus making them beggars.

William Lee went home without his patent and women continued hand-knitting.

But does new technology really kill jobs?

The Industrial Revolution did destroy some traditional jobs, but it also created new ones. Without the technology of the Industrial Revolution, we could never have the standard of living we have today. In the short term, it was a difficult transition; in the long term, it was very worth it.

“Short term pain for long term gain” is likely to be true of artificial intelligence too. Our generation will go through a difficult transition so that future generations can enjoy a higher standard of living.

Here’s what we know about the future of work:

“We are being afflicted with a new disease of which some readers may not have heard the name, but of which they will hear a great deal in the years to come — namely, technological unemployment” — John Maynard Keynes, 1930

In 2013, Carl Benedikt Frey and Michael Osborne predicted that 47% of jobs can be easily automated in the next couple of decades.

They categorized jobs using a 2x2 matrix created by Autor et al. in 2003. They labelled jobs as either routine or nonroutine and either analytic/interactive or manual.

Four Categories of Workplace Tasks by Autor et al., 2003
  • Analytic, or cognitive, jobs use your brain. They involve working with numbers, information, or technology. Engineers, accountants, chefs, and astronauts all have analytic jobs. Analytic jobs require highly-skilled workers with a good education.
  • Interactive jobs require good social skills. They involve working with people. Teachers, nurses, receptionists, and YouTube personalities all have interactive jobs. Interactive jobs require excellent communication skills. They’re included with cognitive jobs because both require a lot of skill and have the potential to pay well.
  • Manual jobs involve working with your hands. Kitchen prep, cleaning, landscaping, truck-driving and painting houses are all considered manual jobs. When a job requires creativity, like being a chef or a landscape designer, it is considered an analytic job. Manual jobs don’t usually require much education.
  • Routine jobs cover the same tasks every day, or month, or year. Loading boxes into a truck is a routine job; so is entering data into a spreadsheet or adding up numbers. Routine jobs are more easily automated.
  • Nonroutine jobs are always changing. Preschool teachers constantly have to react to their students’ unexpected behaviours. Advertisers need to think of new and creative ways to reach their target audience. Janitors need to respond to unexpected cleaning tasks or damage. Tasks that require creative responses to new things are less likely to be automated.

In 2003, Autor et al. concluded that routine tasks would be automated and nonroutine tasks would not. Computers need rules and routines, they thought. Nonroutine tasks are safe from automation.

“[T]he unbiased decision making of an algorithm represents a comparative advantage over human operators” (Frey and Osborne 2013)

In 2013, Osborne and Frey challenged this assumption. Computers don’t always need detailed rules and routines anymore. They’re learning to drive trucks, diagnose diseases, and identify faces. Nonroutine tasks can be automated too.

Osborne and Frey actually argue that machines will become better than humans at tasks like fraud detection or medical diagnoses, because they’re less biased.

Occupations that require subtle judgement are also increasingly susceptible to computerisation. To many such tasks, the unbiased decision making of an algorithm represents a comparative advantage over human operators.

In the past, new technologies have replaced low-skilled jobs. Now, Frey and Osborne claim, they could replace knowledge workers too.

Frey and Osborne determined that 47% of jobs could be automated in the near future — within the next couple of decades. Routine tasks were likely to be automated, as well as tasks requiring unbiased judgment.

Although Frey and Osborne determined that many jobs could be automated, they didn’t consider whether or not they would be automated.

Automation can be expensive, and workers can be cheap. Just because your company can automate your job, doesn’t mean they will. It may well be more profitable to continue to pay your salary!

Technology to automate marking of multiple choice questions has existed for decades, but many teachers still mark exams by hand. Schools don’t usually want to invest in expensive labour-saving devices. It’s often cheaper or more convenient for schools to have their already-salaried staff mark exams.

You can expect the same from artificial intelligence. The technology will probably exist to replace at least some parts of your job, but your boss might not choose to buy it!

“Labor will become less and less important. . . More and more workers will be replaced by machines. I do not see that new industries can employ everybody who wants a job.” — Wassily Leontief, 1952

YouTube personality Saphira Howell talks about her job creating online content.

Some academics, like Frey and Osborne, have attempted to calculate which jobs can be replaced. Other academics, like Daron Acemoglu and Pascual Restrepo, are calculating how many jobs will be replaced.

Acemoglu and Restrepo are using economic models to estimate the impact of artificial intelligence on society. They write:

As long as the rate of automation of jobs by machines and the creation of new complex tasks for workers are balanced, there will be no major labour market decline.

Like Frey and Osborne, they expect AI to accomplish more and more routine tasks. This trend will free up people and money to pursue new careers.

Think of all the new jobs that have emerged over the last two decades . People today create online content for a living, or work for Facebook, Amazon, or Google. These jobs were unimaginable 30 years ago.

Additionally, artificial intelligence is expected to increase productivity and government revenues. Governments will have more money to invest in social services, health care, and education. They could hire more highly-educated teachers with small class sizes or mental health professionals for everyone who needs help.

Acemoglu and Restrepo predict that new jobs will be more accessible to highly skilled workers. They write that “new tasks favor high-skill workers who tend to have a comparative advantage in complex tasks.” They expect that new jobs will complex and nonroutine, requiring skills and education, at least at the beginning.

However, they expect that jobs will become more standardized and easier to learn over time. Acemoglu and Restrepo write, “Because new tasks become standardized, they can over time be as productively used by low-skill workers.” Think about an SEO Expert — fifteen years ago, it was a brand new job, with few mentors or teachers. If you wanted to work in SEO, you needed to teach yourself. Now, there are plenty of standardized options for learning SEO. You don’t need to be a leader in the field in order to get a job; you can take a course or work for a company that will train you.

To be on the leading edge of new jobs, you’ll need skills, education, and creativity. But over time, jobs for less skilled and less educated workers will reemerge.

As machines continue to invade society, duplicating greater and greater numbers of social tasks, it is human labour itself — at least, as we now think of ‘labour’ — that is gradually rendered redundant.” — Robert Heilbroner, 1965

In Acemoglu and Restrepo’s model, artificial intelligence will increase inequality between the highly skilled and less skilled. The highly skilled will be able to take advantage of complex new opportunities, while the less skilled need time to adapt.

Other economists forecast another kind of inequality: AI will increase the gap between owner and worker.

Imagine you own a bakery. You employ staff to help you bake bread, answer customer’s questions, and complete deliveries. Half of your profit goes to paying their wages.

Over the next several years, you are able to buy automated breadmakers. You begin taking all orders online, answering common customer questions using a chatbot. Robots load bread onto a driverless truck, which completes deliveries for you.

All of these innovations cost money, but they mean you no longer have to pay for staff. At first, the money you save from paying for staff goes to paying off your new robots, but over time it’s just money saved. You’ll have more money in your pocket every night than you did before. Your former employees won’t.

In the future, owning your own business and leveraging AI to increase profits will be more important. Owners will become increasingly powerful, unless the government intervenes.

Routine jobs and jobs that require unbiased judgment will increasingly automated over the next few decades. Highly skilled workers will retrain for exciting, new opportunities. Owners will enjoy increasing profit margins. Governments and society as a whole will become richer, allowing more investment into social goods like education or healthcare.

The best way to prepare for the future is to get comfortable learning new skills and taking new opportunities. Repetitive, boring tasks will be automated and replaced with new, creative, and challenging jobs.

For those who can retrain quickly and are comfortable trying new things, artificial intelligence means new opportunities.

Special thanks to Karen Jeffrey, doctoral researcher at King’s College London, for her thoughts on automation and the labour market.

The Startup

Medium's largest active publication, followed by +565K people. Follow to join our community.

Kirsten Horton

Written by

Canadian-born Londoner. I try to leave things better than I found them.

The Startup

Medium's largest active publication, followed by +565K people. Follow to join our community.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade