Don’t Fear The Rise Of The Robots
Why AI-Based Automation Won’t Lead To Mass Unemployment
Since the dawn of the industrial age, a recurrent fear has been that technological change will spawn mass unemployment. By and large, economists’ prediction that people would find other jobs, though possibly after a long period of painful adjustment, has been proven correct.
— Economist Kenneth Rogoff
The fear of automation is as old as automation itself. One of the earliest and best-known examples of this fear occurred during the Industrial Revolution of the early 1800s. Power looms, a recent innovation at the time, were coming into widespread use. These new machines largely automated the cloth making process, which meant that handloom jobs would soon become all but obsolete. Fearing their livelihoods are at stake, enraged textile workers did the only thing they thought might save them: They destroyed power looms, burned factories, and killed factory owners. Newspapers infamously called them “Luddites”, after Ned Ludd, an anti-technology crusader of the time.
Ultimately, however, the Luddites’ violence did little to stop power looms from displacing handlooms. Yet contrary to initial concerns, textile-related jobs skyrocketed between 1830 and 1900. The reason for this came down to simple economics. Lower manufacturing costs meant lower prices for consumers; lower prices meant sharply increased demand for cloth; and greater demand for cloth increased demand for machinery mechanics, maintenance workers, and overseers — respectably-paid jobs that many Luddites, ironically enough, ended up in.
In fact, this phenomenon of work “taken” by machines being replaced by new and different jobs has played out many times throughout history. Cars replaced horses, which meant horseshoe makers and buggy builders were out of work, but new jobs, such as car assemblers and gas station attendants, were created in their place. Computers reduced the need for secretaries and clerical workers but created demand for programmers and IT specialists. The internet replaced travel agents, librarians, and retail salespeople with online travel agencies, search engines, and e-commerce— technologies which require a huge workforce to build and maintain.
However, many say this time is different, that technological progress will no longer have a positive impact on labor demand. As evidence, they point to the recent development of artificial intelligence or AI as it’s commonly called. Their main argument goes something like this: Past automation (like the steam engine and the power loom) primarily replaced human muscle, whereas AI (like autonomous vehicles and robots) will replace the human mind, hence making this new technology vastly more disruptive than anything before it.
On the surface, this argument makes a lot of sense. After all, what jobs would be left for humans to do in a world dominated by hyperintelligent machines? A closer look, however, reveals at least three problems, each of which suggests that the future of human employment won’t be anywhere near as bad as feared:
- Today’s AI is a collection of fairly primitive pattern recognition algorithms. While effective on simple tasks, like image recognition, these algorithms fail on complex tasks that require common sense, intuition, and adaptability. Take driving as a case in point. It’s supposedly on the verge of being automated away, which will put millions of truckers and Uber drivers out work. Yet despite years of development, Tesla’s Autopilot — arguably the world’s most advanced self-driving system — still can’t even safely change lanes on a highway, let alone navigate congested city streets and tight parking spaces. Human-level autonomous driving won’t happen overnight; rather, like all technological advances, will take at least many decades of slow, painful, and incremental research.
- Since it’s easier to see the jobs we’ll lose than the ones we’ll gain, employment predictions tend to be skewed toward negative forecasts. But it’s always been that way. Imagine telling someone a century ago that their great-grandkids would be app developers or SEO specialists or social media managers. Nobody back then could have predicted that these jobs would exist. Similarly, those of us alive today are clueless about what kind of employment opportunities will be created going forward. But this unpredictability doesn’t mean that human workers are doomed.
- People adapt to automation. Bank tellers, for example, used to spend the bulk of their workday handling cash withdrawals. Once ATMs eliminated the need for that tedious task, tellers could focus on the more human elements of their profession, like providing personalized customer service. They needed to adapt their skill sets, but they didn’t lose their jobs. The AI revolution will be more of the same. Structured, repetitive jobs — like accounting — can be outsourced to a machine. But jobs that require curiosity, creativity, and/or empathy — treating patients, for example — cannot. Rather than disappear, jobs will evolve, and as productivity booms, so too will demand for uniquely human labor.
Before concluding, it’s worth acknowledging one small but important drawback of automation: Some workers will inevitably be displaced by it. Those in easily automatable jobs are obviously most at risk. Thankfully, a potential solution to this problem already exists. It’s called Trade Adjustment Assistance (TAA), a program that provides financial assistance, retraining, and reemployment services to those who lose their jobs due to trade with foreign countries. Extending TAA benefits to workers who lose jobs due to automation, as some politicians have proposed, can ease their transition to the AI-automated economy.
That said, any temporary inconvenience of this transition is a small price to pay given the immense long-term technological, financial, and societal benefits we’ll gain. From the power loom to the robotized modern factory, automation has increasingly freed us from repetitive and laborious tasks, allowing our human creativity and intelligence to flourish. It’s made the world a better place and will continue to do so.