3 ms·

AI is not eliminating jobs

6 points by tim_sw 8y ago

notus 8y ago
Wow that article was.... a little crazy? Very disorganized and making a lot of weird assumptions.
nabla9 8y ago
That's not what those actually know something say. Also, the AI eliminating jobs is the wrong question. The right question is the relation between automation (and AI) and wages. Automation and labor income have clearly very complex dynamics. It can go both ways. It's the interaction of multiple factors that determines what is the outcome. Listing just few complexities: Short term reduction in the labor compensation seems possible and there is some evidence supporting it. When robot manipulators and vision gets over some threshold where they become almost as agile and smart as human hands, it will cause major disruption. On the other hand, even in the worst case human labor will find areas where it has comparative advantage even if the human labor more expensive and provides lower quality is absolute sense. But what happens to labor compensation is different thing. Labour augmenting automation is different from labor replacing and they both exist. Labour augmentation can increase wages and labor productivity. Automation can also lower incomes and increase incomes in different income brackets. Automation is capital investment. If increases capital intensity of products enough it can reduce wages and demand.
- projektfu 8y ago
  There are two issues that economists, such as Dean Baker, have identified with the AI story. The first is that productivity growth has slowed substantially. We were increasing productivity much faster in the era of labor intensive extraction, manufacturing and farming than we are now in the era of services and knowledge work. What the AI people are predicting is some kind of rapid, nonlinear productivity black swan event. We creep along at 1.5% productivity growth year after year until all of a sudden we see 10% per month putting tons of people out of work and on the dole. That's fine for a worst case but black swans are, by their nature, unpredictable. The second problem is that productivity growth has generally been good for wages. A huge concern in the US in 1982 was that productivity growth had appeared be 0 since 1978 with periods of declining productivity canceling out any increases. The problem was that wages would stagnate. At the time, it was probably caused by low investment due to high interest rates, but there was a lot of handwringing about American manufacturing and whether it could remain competitive with Japan, which was experiencing increasing productivity at the time. We can look at specific examples of AI failing to realize their goals, but if the overall story of AI was one of replacing workers, it would be shown in the productivity numbers. People who predict a disruptive AI event need a lot more than the spectre of all the trucking jobs disappearing overnight.
crb002 8y ago
Machine vision is. Cash input to self checkout kiosks has now surpassed the efficiency of waiting in line by accepting bills at much higher accuracy.

6 points by tim_sw 8y ago

Wow that article was.... a little crazy? Very disorganized and making a lot of weird assumptions.

That's not what those actually know something say. Also, the AI eliminating jobs is the wrong question. The right question is the relation between automation (and AI) and wages. Automation and labor income have clearly very complex dynamics. It can go both ways. It's the interaction of multiple factors that determines what is the outcome. Listing just few complexities: Short term reduction in the labor compensation seems possible and there is some evidence supporting it. When robot manipulators and vision gets over some threshold where they become almost as agile and smart as human hands, it will cause major disruption. On the other hand, even in the worst case human labor will find areas where it has comparative advantage even if the human labor more expensive and provides lower quality is absolute sense. But what happens to labor compensation is different thing. Labour augmenting automation is different from labor replacing and they both exist. Labour augmentation can increase wages and labor productivity. Automation can also lower incomes and increase incomes in different income brackets. Automation is capital investment. If increases capital intensity of products enough it can reduce wages and demand.

projektfu 8y ago
There are two issues that economists, such as Dean Baker, have identified with the AI story. The first is that productivity growth has slowed substantially. We were increasing productivity much faster in the era of labor intensive extraction, manufacturing and farming than we are now in the era of services and knowledge work. What the AI people are predicting is some kind of rapid, nonlinear productivity black swan event. We creep along at 1.5% productivity growth year after year until all of a sudden we see 10% per month putting tons of people out of work and on the dole. That's fine for a worst case but black swans are, by their nature, unpredictable. The second problem is that productivity growth has generally been good for wages. A huge concern in the US in 1982 was that productivity growth had appeared be 0 since 1978 with periods of declining productivity canceling out any increases. The problem was that wages would stagnate. At the time, it was probably caused by low investment due to high interest rates, but there was a lot of handwringing about American manufacturing and whether it could remain competitive with Japan, which was experiencing increasing productivity at the time. We can look at specific examples of AI failing to realize their goals, but if the overall story of AI was one of replacing workers, it would be shown in the productivity numbers. People who predict a disruptive AI event need a lot more than the spectre of all the trucking jobs disappearing overnight.

crb002 8y ago

Machine vision is. Cash input to self checkout kiosks has now surpassed the efficiency of waiting in line by accepting bills at much higher accuracy.