Automation isn't new—we've seen the effects of automation due to technologies such as the Gutenberg printing press and the steam engine in the past. Despite the rampant fears that these technologies would lead to mass displacement of workers, it turned out that the fears were largely unfounded. The displacement was much slower than anticipated, and the new technologies gave rise to a large number of new jobs.
When looking back at history, it's easy to brush off the fears of AI being equally meaningless, but there are a few important reasons why it's different this time around.
👷 AI Replaces "Human Skills"
Automation used to mean teaching robots to follow programmed instructions. We could create automations to handle tasks that we didn't want to do but knew how to do. That left humans with jobs that we considered more "complex"—tasks that we didn't think were automatable. AI enables us to replace those same skills that we considered innately human, such as responding to human language or recognizing movement.
This logic may echo the fearmongers back in the early days of the printing press, but that would be comparing apples to oranges. AI is encroaching on what we considered to be the role of humans in the workforce in a way that doesn't remotely resemble what we've ever seen before.
Human jobs in the age of AI will likely be significantly more scarce and, on average, will require a much higher level of skill than our job market is currently equipped for. AI isn't technology automating the "boring work"—it is automating a large chunk of the work that labor workers, like call center representatives or truck drivers, rely on for a living.
And it's not only labor workers that are in the line of fire—white-collar workers like paralegals are at the risk of losing their jobs to machines as well.
🧠 AI Creates Its Own Logic
Automation using AI means that we can program more tasks that we know how to do and that we can program tasks that we don't know how to do. Unlike automation technologies of the past, AI creates its own decision-making processes, meaning that it can solve tasks in a way that we may not comprehend.
For example, when the world's best Go players competed in a game of Go against the AI-powered AlphaGo computer, several described its decision-making process as "alien" or "from an alternate dimension." AlphaGo repeatedly made decisions during the game that perplexed the leading human experts in the game. And it repeatedly won.
AI's ability to find better ways to solve a problem than its human counterparts increasingly threatens the role of knowledge workers in the future of work—something most people would have thought was unimaginable just a few decades ago.
🏎 AI Is Developing Very, Very Quickly
As a technology, AI is progressing at breakneck speeds, and its progression is accelerating. To highlight this, here are a couple of key takeaways from the 2020 AI Index Report.
From a research and development perspective, we're seeing an acceleration in publication outputs across all key players, including the United States, China, and the European Union. The total number of publications grew by nearly 12 times between 2000 and 2019 (an annual increase of 34.5% in the last year alone). Over the same period, the percentage of peer-reviewed publications increased from 0.82% of all publications in 2000 to 3.8% in 2019.
From a performance perspective, we're seeing massive improvements in both computational power and model performance. As an example, in the field of computer vision, "the time it takes to spin up a model that can classify pictures at state-of-the-art accuracy fell "from around one hour to around 4 minutes" in just 18 months—a 16x jump in training speed. Image segmentation (a computer vision capability that separates an image background and its subjects) has increased in precision by 72% in the last three years.
When looking at past technological revolutions, it's easy to dismiss concerns of mass displacement resulting from AI as unfounded. However, there's enough reason to believe that this time will, in fact, be different.
On a personal note: Contrary to what it might seem, I didn't write this article to fearmonger. I firmly believe that the first step to creating the future that we'd like to create for ourselves is recognizing the path that we are on; this equips us to take the steps we need to create the future we want for ourselves and our society. The AI revolution doesn't need to spell doom and gloom—we can make it work for all of us. The first step, though, is becoming aware. So thank you for taking the time to become more aware.