I, For One, Welcome our Robot Overlords | Opinion

A robot developed by automation parts maker Omron picks an item from a shelf during a demonstration at the World Robot Summit in Tokyo on October 18, 2018. BEHROUZ MEHRI/AFP/Getty Images

For the last 17 years, the Japanese robotics company FANUC has been operating a very special factory. At this facility, robots build other robots—50 of them each day. The building is "lights out," meaning it operates without human intervention. It can do so for a month at a time.

If this sounds like the beginning of a Philip K. Dick short story, it is. (Thankfully, though, FANUC employees get along far better with their robots than Dick's protagonists do.)

Manufacturing isn't the only sector where firms aspire to the lights-out model. Recently, a packing and shipping warehouse for Chinese retailer JD.com made headlines by going almost entirely robotic. And with the prospect of automating white-collar work growing ever more likely, we could see lights-out offices one day, too. The trend has many concerned about mass unemployment. Though estimates vary, most experts say that America will lose tens of millions of jobs to automation in the next decade.

However, rumors of the death of jobs are premature. Here's why: Going lights-out doesn't necessarily make a facility more productive, even in manufacturing. Consider Toyota's 1970s thwarted attempts to fully automate its plant in Tahara, Japan, or Tesla's more recent failure to deliver Model 3s using a highly-automated production line.

There's a reason that automation hype is outpacing reality: Automation transforms markets just as easily as it takes jobs. And the same developments that have affected manufacturing automation are likely to impact any other industry that attempts to create a "lights-out"
model. Even in a highly automated world, humans will remain very necessary.

The Fight for the Factory of the Future

When we imagine the factory of the future, we tend to picture a more efficient, cost-effective version of the factory of the past. But that's wrong. The technologies that power industry 4.0 are also changing the ways factories function in today's hyper-competitive global economy. Factories aren't just becoming more automated—they're also becoming more flexible. And human beings, not robots, are the key to that flexibility.

How? Here's an example. While automation threatens low-skilled manufacturing jobs, it also drives down costs and reduces barriers to entry for newer, smaller manufacturers. By putting out small batches of quickly-changing products, these new entrants drive up the number of choices available to consumers. The beer industry is a great example. Twenty years ago, macro brews like Budweiser and Miller dominated; today, smaller craft breweries take up nearly a quarter of the market.

Over the long term, these trends create feedback loops that shift demand patterns. Consumers come to expect greater variety and a constant influx of new products. As a result, more established players must offer more choice, further reinforcing the shift in consumer expectations. Under these circumstances, scale can become a liability; after all, it's harder for large factories to change what they produce on a dime. This is part of the reason we see major beverage companies investing more in production facilities for their craft beer brands—small and flexible is the future.

What does this have to do with the lights-out factory? Large factories become especially inflexible when they're stocked with robots specialized for particular tasks. After all, more variety means more opportunity for breakdowns and disconnects—problems that people are better-suited to fix.

Lights-out factories have run into these issues from the beginning, with no obvious solution in sight. For example, early Apple manufacturing facilities were much closer to lights-out than the factories that assemble iPhones today. At one Macintosh plant opened in 1984, nearly 90 percent of the assembly process was automated. It closed only six years later. Steve Jobs ran into the same problem when he attempted to launch the NeXT computer in 1988. His highly automated factories were unable to adapt when demand for the new devices flagged. Today, Apple relies on humans to assemble its products.

Where machines bring predictability and consistency to production processes, people bring more adaptive decision-making skills. And our human capacity to quickly respond to varying inputs is crucial, because some tasks cannot be automated without reducing the overall quality and safety of the output. Most manufacturers know this, though they don't necessarily publicize it. For example, BMW is famous for automating its factories, but there are still more than 10,000 employees at its South Carolina manufacturing campus.

The future of the workforce

What does all of this mean for industries outside of manufacturing? It means that despite the hype around cashierless Amazon Go stores and bookkeeperless back offices, it's unlikely we'll be seeing convenience stores or workplaces go totally "lights-out" anytime soon. Machine learning models, which must be trained and tuned to specific sets of data, are still less flexible than human employees—an important caveat in the constantly changing contemporary economy. Other industries will run into the same hiccups as manufacturing when they try to automate at scale.

The savviest manufacturing operations typically put people to work alongside collaborative robots, or "cobots." Unlike the clunky machinery of the 20th century, cobots are designed to share space and interact with humans. According to an MIT study, cobots at a BMW plant are 85 percent more efficient than either humans or robots working alone.

Most important to the factory of the future, however, is not the way humans guide robots, but the way that technology can guide and empower humans. Tomorrow's "connected worker" technologies have the potential to empower human actors with IoT-enabled capabilities, allowing workers to gather data continuously and interact with the factory's automated systems more smoothly. Human and machine will work as one.

This alternative to going "lights-out" is revolutionizing more than just the manufacturing industry. The city of Baltimore recently announced plans to install "smart trash cans" that will notify sanitation workers when they need to be emptied, increasing collection efficiency. In Amazon's warehouses, which employ more than 125,000 people across the US, workers oversee robots that perform the tedious task of stacking and moving boxes, freeing people up for more high-level work. In each of these cases, automation augments human creativity, dexterity, and thought by adding a guardrail of consistency and predictability.

Concerns about automation and unemployment are real, and they deserve to be taken seriously. However, the news from the front lines of the automation revolution suggests that the workplaces of the future could be friendlier to humans than many people imagine. If other industries follow manufacturing's lead, we'll see tomorrow's advanced machines collaborating with, rather than replacing, their human coworkers.

Lawrence Whittle is the CEO of Parsable.

The views expressed in this article are the author's own.​​​​