We Are Not Prepared For The Rise Of Automation

Photo Via Brookings Institution

ANTONIN GROSCLAUDE-EVANS: Though we are accustomed to the notion of robots attacking our government in science-fiction movies, this threat is materializing in our real world. In 2019, AI hit squads from Harvard College overran Idaho’s official Medicaid application website. Overworked employees couldn’t differentiate between human and AI-generated requests, leading to thousands of taxpayer dollars wasted…

While the claim that this is another step in AI’s global domination plan is hyperbolic, the underlying threat is very real: that of AI and automation. Throughout its history, humanity has strived to improve productivity through technology. 20 accountants with abacuses are as productive as, say, five accountants with calculators, two accountants using Excel or one accountant with a dedicated computer program. This raises a question: what happens when we go from 20 accountants, to five, to two, to one and finally to zero? To tell the truth, there is very little time left before very big problems appear.

Professors Carl Benedikt Frey and Michael Osborne of Oxford University have created a methodology to answer the following question: how susceptible to automation will jobs in the U.S. be by 2033? The findings they share are quite concerning: “According to our estimates, about 47 percent of total U.S. employment is at risk.” That is a predicted unemployment rate almost double that of the Great Depression.

Critics have attempted to minimize automation’s threat, with some claiming that organized labor can protect jobs from automation. However, history has proven otherwise. In 1945, driverless elevators were slowly appearing in New York City. Many were initially skeptical of the idea of a machine transporting them to their destination, comparable to how the public thinks about driverless cars today. Still, driverless elevators were soon becoming popular enough to threaten NYC’s 15,000 elevator operators, who went on strike in protest. In response, skyscraper’s elevators were automated despite public concerns, and elevator operators are practically non-existent today. The parallels between 1945 New York and automation’s current stakes cannot be ignored.

Though proponents of automation have argued that it will create better paying jobs,  that is also incorrect. As of September 2022, 39 of the 40 most common U.S. professions also existed in 1800, with the sole exception being “Computer Occupations.” Even though new professions are being created, they generally employ far fewer people than professions being automated. For example, in 1979, General Motors employed 618,000 Americans for a $281.8 billion revenue (adjusted for inflation). Today, Google has over 4 times fewer U.S. employees, 140,000, for the near-identical revenue of $280.8 billion.

In the face of the threat of automation, many have suggested that it will leave high-skill jobs largely intact. Unfortunately, that is also untrue. AI is already being trained for jobs like medical care, analysis, stock brokerage and accounting. Even jobs we cannot fathom being automated are at risk. Take lawyers, for example: it seems nonsensical to picture a robot delivering speeches to juries in court. However, most of the work involved in the legal profession consists of reviewing court cases and various documents — tasks AI does immeasurably faster than any human could. It’s likely that in the future, lawyers will still appear in court, but legal firms will hire far fewer people.

Is there no solution? Is AI fated to overrun humanity, creating widespread economic misery? Not quite. Despite this pessimistic analysis, actions can be taken to smoothen the transition. In September, professors Frey and Osborne, published an article in The Economist, dissecting how well their 2013 predictions had fared, 10 years after their previously mentioned paper was published. The two generally reasserted the likelihood of the labor market’s widespread automation. However, they highlighted rising political opposition to automation, as seen during the recent screenwriters strike. Congresspeople have introduced bills to protect workers from automation, like Sen. Sherrod Brown’s 2019 bill guaranteeing advance notice, job training and benefits for workers laid off due to automation. Such political action goes a long way in protecting those most vulnerable to automation. 

However, individual nations cannot address the threat of automation alone. One country cannot simply ‘ban automation,’ lest they risk falling behind technologically. To best adapt to automation’s looming consequences, international diplomacy is needed to create a common agreement regulating automation, sharing the burden while protecting workers. Unfortunately, global decision-making bodies are slow-moving, and the international community has been historically slow in addressing other global concerns like climate change.

With just around a decade left to answer the question of automation, there’s no time to lose in working towards a solution. 

Antonin Grosclaude-Evans is a first-year staff writer for On The Record studying in the SFS. He is a Franco-American born in Boston, Massachusetts hoping to study international politics.