Automation has frightened workers for over a century. Ever since the dawn of the industrial revolution, workers have feared machines may replace all human work. Mechanical muscles worked faster, stronger and tirelessly to reduce jobs, and led to efficiently cheap abundance.
However, these machines were dull, towering behemoths that could only perform simple and repetitive tasks in manufacturing or resource production. While many scholars scoff at the idea of not having a job in the future, do not be too quick to mistake this as the same situation. Technology and the internet are now involved in every facet of our daily lives, but that is still a fairly recent change. The Information Age has only been going for a few decades, and we are now noticing terrifying new problems that result from data processing and prediction.
Like my fellow classmates, I was curious about lower job security in the future, even with a college degree. This curiosity led to several educational videos discussing automation, including a CGP Grey short film and an animated Kurzgesagt video. “Humans Need Not Apply,” CGP Grey’s 2014 short film, gave some sobering data that makes the Great Depression preferable to our situation. Whereas the Great Depression only impacted a quarter of the working American population, roughly half of all jobs are likely to become automated.
Kurzgesagt’s video, “Why Automation is Different this Time,” only provides more frustrating data. Machines can improve in increments far faster than humans, swapping code to suit whatever task is at hand. The lower rate of error, increasing rate of speed and shareable updates lead to humans not even being considered for most future positions.
Studies have tested the impact of automation for decades, including a major 2013 study by Carl Benedikt Frey and Michael A. Osborne of the University of Oxford. Aiming to determine the future of employment, this study asked how susceptible jobs are to computerization.
Factoring in the advancements in machine learning, lower computational cost and data collected by respected agencies, the results outline an inevitable exponential curve in efficiency. After testing over 700 occupations, 47 percent of all work is at risk to become automated within the next few decades. The timeline for this impact is still to be determined, but the data all leads to machines becoming far superior to humans. Whereas humans take years to specialize, a computer can be upgraded and duplicated to suit needs of production.
Erin Winick of the MIT Technology Review summarized several scholastic studies last year regarding automation. Her findings were that there is not a clear definition between optimism and devastation. The overall time-frames are skewed based on the technology discussed, leading to figures in diverse occupational fields. Many studies focused on specific industries, including the transportation upgrade with autonomous vehicles. While numerous studies were created by dozens of global experts, none seemed to be “on the same page.” This led to one specific conclusion: there is no clear answer as to how many jobs will actually be lost to automation, but it will happen over time.
I wanted to help calm radical notions that automation addresses, but after my research, I cannot offer much except a warning. Automation is not coming because it has already been around for decades. The difference this time is that our mental performance is now inefficient, as we are limited by our slow bodies and living costs. Perhaps the most infuriating statistic is U.S. workers provided 194 billion hours in both 1998 and in 2013, yet became over 40 percent more productive in that time, according to Winick.
If that was the case, why is the wealth gap getting wider when we should all be working less? Jobs will disappear in faster intervals, leaving us extremely productive but unable to purchase anything if only jobs provide income. I am not sold on universal basic income, but few alternatives will keep our global economy functioning when the majority of humanity is replaced by smarter machines.