Will Robots Take Over Our Jobs?
You've seen the headlines: half of all jobs automated, machines doing most of the work within a few years. This week you get to check them. The single most useful thing you'll learn is that almost no job can be fully automated — and almost every job has parts that can be. Those two facts explain why the scary articles and the reassuring ones can both be telling the truth. You're also making decisions about your own future against these predictions, which makes this less abstract than most topics.
Reading List
Half of these are videos, so this is a lighter reading week than most. They disagree with each other on purpose — one argues this time really is different, another argues it isn't.
Required
- Why 'the Future of AI Is the Future of Work' — MIT Sloan
- The Rise of the Machines — Why Automation Is Different This Time — In a Nutshell (video)
- Don't Fear AI. It Will Lead to Long-Term Job Growth — World Economic Forum
- Will Automation Take Away All Our Jobs? — David Autor (video)
If You Want More
- The Triple Revolution — written in 1964 and startlingly current.
- How to Solve AI's Inequality Problem — MIT Technology Review
- A World Without Work, with Daniel Susskind — Oxford Martin School (video)
One-Week Plan
Here's the shape of the week. Your guide will set the exact timing for your section.
What the Forecasts Actually Say
Write first: will robots take your job? And where did that impression come from — a video, a headline, a conversation?
Then the numbers everyone quotes. An Oxford study of 702 jobs said 47% could be done by machines within twenty years. An OECD paper put 210 million jobs at risk across 32 countries. The World Economic Forum said machines would go from doing 29% of work to over half by 2025.
Then the finding that changes what all of those mean. About half of all workplace activities could be automated with technology that already exists — but fewer than 5% of jobs could be fully automated, while almost every job has some part that could be.
That distinction is the most useful thing in this week. Automating tasks is not the same as eliminating jobs. When you next see a frightening statistic about AI and work, the first question to ask is which of the two it's counting.
Read the Charts Carefully
- Richer countries face less automation risk. Worth asking why — and why two countries with the same income can still differ. (The answer is what industries they have.)
- Most at risk: food preparation, then construction, cleaning, and driving. The obvious pattern is manual work — but sales and customer service are on the list too, so the obvious pattern isn't the whole story.
- What actually gets automated: physical work in predictable settings, processing data, collecting data. Notice the word predictable. A warehouse is predictable; a building site isn't.
- A different measure, a different answer. "AI impact" measures how much AI shows up in a job at all — not whether it replaces the job. On that measure, the most exposed workers are white-collar professionals: radiologists, lawyers. Machines have become extremely good at prediction and pattern-spotting.
And a harder question: even if automation creates more jobs than it destroys, who gets them? The evidence from the US and Europe suggests the benefits flow to people with technical and managerial skills, which could widen the gap rather than close it.
People Have Been Scared of This for 200 Years
In the 1810s, English textile workers smashed the machines they thought would replace them. In 1850 New York tailors struck over sewing machines. In 1860 grain shovelers unionised against grain elevators. In 1890 England required a person to walk in front of every vehicle waving a red flag. In 1930 Keynes gave the fear a name: technological unemployment.
And two times the fear turned out to be exactly backwards:
- The ATM. Everyone expected cash machines to eliminate bank tellers. The number of tellers actually went up between 2000 and 2010 — because ATMs made branches cheaper to run, so banks opened more branches, and each branch still needed people.
- High school. When automation destroyed farm jobs around 1900, America responded by requiring everyone to stay in school until 16. It was hugely expensive. It also turned out to be one of the best investments the country ever made.
But here's the fair challenge: "people were wrong before" doesn't prove they're wrong now. Try arguing both halves — the case that this is the same old panic, and the case that this one is genuinely different.
Before you leave, find a resource to bring on Thursday.
What Work Is For
Share what you found, then move from predicting to deciding.
- Retraining. The World Economic Forum estimates half the world's workers need new or upgraded skills by 2025. Is that realistic? And who should pay for it — companies, governments, or you?
- Guaranteed income. In 1964 a group of thinkers argued that since income comes from jobs and machines are eliminating jobs, we should break the link between income and work entirely. Sixty years old, and it sounds like a current argument.
- Keynes's prediction. In 1930 he said we'd be working 15-hour weeks by 2030 in an "age of leisure and abundance." We're nearly there. Why hasn't it happened?
Finish with the questions worth arguing about: would it be good if machines took the boring and dangerous jobs? What would people do with themselves if work weren't necessary? And should we try to slow automation down at all?
Score Your Own Week
What went well, what to improve, what you learned, and whether anything you believed changed. Then the participation matrix. Be honest with yourself.