Automation vs. humanity: A false binary choice

Why robots seem so inevitable

Before you begin panicking about robots and AI, it’s important to understand why a business would choose to deploy a robot over a person.

Let’s take manufacturing, which is where my background lies, and which is often considered ground zero for automation.

You’ve probably read about the resurgence of American manufacturing. And you’ve also probably read about the labor shortage faced by American manufacturers. It may come as a surprise to people who are worried about automation, but humans remain absolutely essential to manufacturing. In fact, Boston Consulting Group estimates that up to 90% of manufacturing tasks are still performed by humans. The exact number varies within manufacturing industry verticals, but it’s safe to say that humans are the greatest contributor of value in the factory.

Human are, unfortunately, also the factory’s greatest contributor to process variability.

So, today’s manufacturers don’t see the problem as one of having too many humans. The real challenges are high turnover, labor recruiting shortages, long training times and worker safety risks — and the unpredictability that these factors create within the value chain. Manufacturing operates on extremely tight timelines (remember kanban?) and low margins; unpredictability is unwelcome because unprofitability often follows.

But if it were just so easy as to summon a horde of robots to kick all the people out, that would have been done by now. The truth is that, even with robotics and AI advancing at extreme rates, there are some very physical limitations on robot proliferation. I’ve outlined three of them below.

Robots aren’t rabbits; they multiply very slowly

A robotic system is a very complex manufactured good. Ask Elon Musk about how easy it is to scale up production of a complex manufactured good, even when incentivized by staggering demand. It’s not.

The International Federation of Robotics expects the global industrial robot population to increase from 1.8 million in 2016 to 3 million by 2020. Sound like a large number? Well, according to Goldman Sachs Research, there are more than 340 million people working in manufacturing worldwide. So, the robot population is actually growing at a fairly slow rate relative to the perceived demand.

If economists Daron Acemoglu and Pascual Restrepo are correct, each robot replaces 5.6 workers. Their estimation seems large at first glance, but it’s actually less than 2% of the global manufacturing workforce.

The market demands skills that robots don’t have

Henry Ford would have killed to have a few good robots in his factories. His was the ideal use case for robots: a high-volume, low-mix product that rarely changes and has very few variations.

Unfortunately (for robots), today’s manufacturing environments are driving toward the exact opposite trend: from mass production to “mass customization.” Lot sizes of one. Here, the manufacturer strives to deliver flexible and even personalized products without incurring the high unit costs associated with artisanship. The best example comes from Ford: 100 years ago, Henry Ford said, “Any customer can have a car painted any color that he wants so long as it is black.” Today, the Ford F-150 has a staggering 4,147,200 build combinations — or two billion, depending on what you count as a variation.

This kind of environment is precisely where robots struggle. “Machines excel in highly repeatable, high-volume operations,” said Peter Marcotullio, vice president of commercial R&D at SRI International. “Unfortunately for machines, the trend in manufacturing is for mass customization — small production runs, more process variations, constantly changing components — which is very hard to automate because of the intrinsic upfront costs of getting a flexible robotic system tooled and programmed. But it’s very easy for a person to adjust on the fly. People are more flexible and can learn faster than machines.”

Behind every robot is a cadre of … humans

Factories operate on a very delicate cadence, marching to the beat of the takt time. The process engineers know exactly how fast a person is supposed to work, and have designed complex systems to ensure that raw materials and components reach every workstation in exactly the right rhythm to ensure the operator never has to stop and look for parts.

If a robot is to work faster than the person it replaces, then the materials flow around that workstation needs to be rebuilt as well. A factory is a complex system, and any change at one node cascades throughout all the adjacent nodes.

That’s why every robot requires an ecosystem of programmers, process engineers and skilled technicians just to get started, and far more significant redesigns to reorient the process around the robot’s natural advantages. The second-order effects cascade through the supply chain, which requires people to address — who, as I’ve already mentioned, are actually the scarcest resource in any factory.

The real way this movie plays out

Everyone loves a good invasion story. Robots, aliens, body snatchers — that’s what fills movie theaters and sells newspapers (or, in this day and age, attracts clicks).

In reality, I expect the relationship between people, robots, AI and future unknowable technologies to continue the same pattern we’ve seen in 200 years of technological change: Some jobs are made redundant, some jobs are enhanced and many new jobs are created.

There’s no doubt our society will reorient and be reoriented by new technology. But the invading robot armies will remain the subject of fiction for a long time to come. The real story — which will be exciting to historians but less so to moviegoers — is one of humans and machines, not humans versus machines.

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.