Automation - Killer or Creator of Jobs

Almost every day there seems to be a discussion in the press about automation, AI, the IoT and the effect it will have on employment. Today, a press release arrived in my inbox about a survey entitled, "The Potential Consequences of a Highly Automated Construction Industry," that concluded that increasing automation in the construction industry could displace or replace as much as 49% of the America’s blue-collar construction workforce (2.7 million workers) and eliminate nearly 500,000 non-construction jobs by 2057. I don’t know if the summation is correct or not. I’m not going to guess if the survey was biased and in which direction. I’m won’t even try figure how a construction survey on the Midwest of the USA arrived in a European electronics editor’s inbox. But, one thing I am sure of is that there will be enormous changes in the employment market over the next few decades.

As part of my last job I worked closely with the automation companies that hope to create the technologies that will cause the predicted job losses and that technology is impressive, even at this nascent stage of development. Robots are becoming more aware and have progressed from following a line painted on the floor, to calculating the best route to take to a location and safely navigating its way there. In some cases the robots themselves decide what the optimal location will be…well not exactly themselves – it would be more accurate to say, with the help of teraflops of readily available computing power and multiple inputs from machines in devices all over the location.

These advancements have happened in a matter of years rather than decades. As I’ve mentioned previously here, at the SPS Drives exhibition, Infineon was even talking about getting people to develop their own robots. The demonstration on the company’s booth showed how local students had developed a scaffolding robot. It, or something similar, will eventually be responsible for some of those Midwest construction job losses.

It’s not just robots that are learning to get around the place themselves, automobiles are too. Every day I get news of more components certified for autonomous driving, or advances driver assist vehicles. There is a huge race between big auto and big tech to get the first truly autonomous vehicle on the road, and once it starts it won’t stop. It is easy to imagine in future newspaper headlines that human drivers are be selfish and dangerous and they keep the speed limit down and the roads blocked as their reflexes can’t compete with silicon. Every death on the roads will be highlighted as an avoidable tragedy. It won’t be too far in the future before there is a movement to ban humans from the road for our own protection!

But, so far we have only been discussing blue collar workers - the drivers, the builders and the machine operators, but that is only a part of the story. If AI can navigate a vehicle across countries it can do accounts, order parts, manage production plants, design circuits – almost any job - including writing news articles, so I better watch my back too. Even doctors better watch out, I have read recent stories in the press talking about how AI can diagnose x-rays and other medical tests more accurately than doctors. How surgeons aren’t getting enough practice as robots are doing easy tasks in theatre and how robots can even stitch up wounds better. I think that the most ironic thing I’ve read is that AI can create more effective AI than the human programmers that initially created it.

I’m sure there’s a lot of exaggeration in these articles, but it’s easy to forget that technology such as autonomous vehicles were thought to be a long way off even a decade ago, and now there are self-driving vehicles on the roads. The technology is still at an early stage of development, but it is progressing at a tremendous speed. I was actually talking to a friend and ex-colleague yesterday, who is now employed by distribution giant Avnet. He has been working on an automated customer service program that is now in open beta that will help with product selection, fulfilment queries, customer support and even engineering. It sounds futuristic, but if you deal with Avnet, you’ll likely encounter it soon. You can read more about that project here https://enterprise.microsoft.com/en-us/articles/blog/microsoft-in-business/how-avnet-enhances-customer-experiences-with-microsoft-azure/ and in February’s PSD magazine.

So it appears almost everyone’s job is under threat in some way or another - or is it? What jobs will be created for those 3.2 million Midwesterners, and when will they be available? In one economics subreddit I frequent, the answer always seems to be that people always found jobs that didn’t exist during the other three industrial revolutions. Farm workers migrated to the cities. There was no mass unemployment caused by the second or third industrial revolutions. The computers that enabled the third revolution even created more jobs than those they displaced. The posters assume that the same will happen this time. We just can’t conceive them at the moment as the pressure for them has not materialised yet, and won’t until AI and the IoT is more advanced. I’m not an expert and I’m not going to guess. I’m just pretty confident that the world in twenty tears time will be a lot different to the one we are living in now.

If you are interested in reading the initial survey that provoked this blog/rant/brain dump, you can find it here https://midwestepi.files.wordpress.com/2018/01/the-economic-consequences-of-a-highly-automated-construction-industry-final.pdf

PSD