The idea in a nutshell : Artificial Intelligence (AI), appears to be different from the incremental technologies (including information technology) we have seen since the release of the jet engine. The productivity improvements involved with AI could ripple through the whole economy and affect every vertical. This is change on an order of magnitude that most people alive haven’t experienced and it makes the potential effects of AI even more dangerous. One fix would be STEM training for displaced workers, especially low skill, low pay men. But there doesn’t seem to be much of that going on.

Evidence in support of the claim

Robert Gordon the famous economist claimed in 2012 that economic growth was over. I read the white paper (it is one of the few bits of internet content that I have ever paid for) in 2013 and it was a life-changer.

In it, Gordon claimed that we face a series of ‘economic headwinds’, including enormous government debts to repay after the global financial crisis. He believed innovation would not save us from that fate.

A central theme to the concepts he outlines is that previous economic growth was fuelled by inventions and progress which affected the whole economy. Examples included electricity, indoor plumbing, air conditioning, rail systems, road systems and the internal combustion engine.

These new inventions, he claimed, released in the 20th century, affected the whole economy and drove global economic growth at a spectacular pace. Since the jet engine, however, ideas and innovations which have been released to markets affect a component of the economy, not all of it. He specifically covered what may appear to be an exception – IT, Information technology and suggested that it had not had the effect that most people believe. Gordon believes that growth since the ‘post war boom’ (including all of the advances made in IT) was part of a period of ‘tepid’ growth.

But did he miss the coming of age and speed of approach of AI ?

Things have changed a bit since 2012. We hear a good deal more about AI than we used to back then. And, ironically, given the thrust of his messaging, AI seems to apply itself to every industry in at least some ways. Just like the benefits of electricity.

One of the other things is the adoption rate for new technology. Below, product adoption is speeding up ‘across the board’. AI will be adopted faster too, in every industry.

chart4

The Tinancial Times and AFR agree

The Financial Times says AI will affect everyone: AI technology will ‘soon replace human workers in all sections of the economy’

The AFR points to:  Mass unemployment and huge increases in share prices could go hand in hand, accelerating and exacerbating income inequality.

Why this might lead to broad wage deflation

When a whole tract of people are made unemployed in an industry those that can apply themselves to the task of working in the next best thing they are good at. You might be a nurse who also is also good at baking cakes. If your job as a nurse is replaced by a piece of technology – for example, some AI, you can then examine your options. You still need to work so you might start to sell your cakes for money. That would increase the supply of bakers and reduce the price of their labour (i.e. reduce the wages of bakers)

The problem is, because AI is going to affect the whole economy, such swathes of people could be made unemployed that it could have the effect of lowering wages for everyone.

A massive retrain is one alternative

One solution to the whole economy effects of AI might be to train displaced people up. It might make sense to do that in the areas which will benefit from AI. This is another change – so called STEM Training requirements of Science Technology, Engineering and Maths. The Australian government has talked a lot and actually done something about STEM training in schools.

Surviving (and I mean the country surviving) the slew of change associated with the dissemination of Artificial Intelligence will require a good deal more investment in this area. We need to support the government in tracking and being aware of people who have been displaced by it’s implementation. So far, there does not appear to have been any investment in re-training programmes by the government for STEM skills.

The CSIRO in their amazingly good report ‘Tomorrow’s Digitally Enabled Workforce’ could not have made their points clearer – unless, perhaps, by providing an estimated likelihood for each of the 4 futures they envisaged. Unfortunately, as it clearly does report – whilst participation in schools in STEM studies is now decreasing.

Summing up

Gordon’s whole point is that most of the productivity and quality of life enhancements that the human race has seen have been around for 60 years or more. Most people take them for granted in the way that Millenials take access to the internet for granted.

Interestingly, he does not appear to have considered Artificial Intelligence in his forecasts. He did address IT (of which, some would argue AI is a part) as a potential positive influence on growth. He discounted it. His point about new technologies not affecting the entire economy is a good one but he appears to have missed an IT based change that was just over the horizon and which invalidates his point. I think AI will have a significant effect on growth in some parts of the economy.

When a whole tract of people are made unemployed in an industry those that can apply themselves to the task of working in the next thing they are good at. This can have the effect of causing a general decrease in the price of time sold to companies (wages.) This is a self exacerbating problem which accelerates the income inequality we’ve already been seeing ignored by government policy and about which most people are in the dark. A huge focus on retraining and STEM, especially for those in later life, is a key step we do not seem to be talking about nearly enough.

Why can’t we use advanced data modelling to determine both the jobs which will be required and work to train especially low skilled men (who will be disproportionally affected) to fill them.

The concepts of AI are hard enough to understand on their own, let alone to imagine at a macro level. Most of us take jet travel, indoor plumbing and cars for granted. We’ve grown up with them. Seeing the world change in the same, fundamental ways that those new rollouts impacted the world is unimaginable for us.

Advertisements