Friday, 24 May 2019

AI and the sky's a-falling 1

Jacobs (1890, p. 182)
Gosh, we humans are slow to recognise patterns. Once we all worked at home doing the menial labour on our tiny landholdings or doing menial labour on someone else's land. When the industrial revolution began, everyone was going to be out of work. The opposite happened: we realised that we could all go and work in a factory, and have some resources left over. We started getting ideas above our station.

So we move on to the computer age, and suddenly we were all going to be out of work again. But instead, there was more work, for even more people. Ditto for automation. 

Now many people are calling "the sky's a-falling" again (Jacobs, 1890, p. 182), this time about AI. Oh, yes, but it is different this time. It will happen faster. We won't have time to adapt (Rayome, 24 January 2019).

Call me cynical, but I remember the 'paperless' office that was going to revolutionise the workplace. Didn't happen. Still isn't happening (although I am largely paperless, myself, very few people or organisations are). I remember how automation was going to swamp us, that all our jobs could be done by programming. Everything would go to via an automated call centre. End of the world. No work for anyone. Didn't happen. Still isn't happening.

Instead we have gone to smartphones and have lots of people now manufacture apps. The once terribly complex computer language and logic has been simplified. Fewer errors happen. We continue to find more and more things for people to do, to engage with, to be challenged by, and to earn a living at.

Yes, some jobs will disappear, but many more will be created. This is utterly unverified, but I read somewhere that once there were something like 200 professions, and now it is more like 200,000 (actually, if any of you have any reliable sources and numbers for this, I would be very interested in hearing!).

There are also some interesting population trends. As soon as we earn over $10 per day, and child mortality falls in line with WHO guidelines, we stop having more than two children (see the Gapminder Foundation for more info). Then we need to consider the dependency ratio (Grant, 2018). This is the number of people who are not in work - retirees, those in education, at-home parents - who need to be supported by a decreasing ratio to working age people, as we continue to live much longer than the three years of paid support intended by global governments as a reward for our life-long efforts. With an average death rate of around 80 years, that is 15 years which governments cover now, rather than 3. A major fiscal blowout. In China they call this "4-2-1" meaning that one working child supports their just retired parents and their four long-retired grandparents (Goldstein & Goldstein, 2015). There is talk about this becoming "8-4-2-1" as life expectancy continues to increase.

So, I am dubious about claims that the world will end because of AI. There are some economists who agree. Take Rainer Strack for example. He has an impressive pedigree in HR consulting with the Boston Consulting Group and is predicting a fairly significant workplace shortage in Germany by 2030 (Strack, 3 December 2014). 

Denning (29 October 2014) has research evidence indicating that new roles are created by private sector organisations which are less than five years old, holding roughly the same pattern since the 1980s. Even more interesting is that those established organisations which are "more than five years old destroyed more jobs than they created in all but eight of those years".

So maturing companies automate. New ones hire people. There will not be enough people for the jobs we currently have. 

Let's grow some entrepreneurs and worry less about whether the sky's a-falling.



No comments :

Post a Comment

Thanks for your feedback. The elves will post it shortly.