My Job Has Finally Been Taken Over By Robots

It has been predicted by a number of people throughout history that new waves of automation will result in widespread job losses and convulsive disruptions. These predictions have proven to be false for centuries.

‍

In the debate over artificial intelligence, the weight of history says that it cannot be different from what we have seen before. The revolutionary nature of ChatGPT has made us rethink our position.

‍

Since the beginning of time, artificial intelligence has been ingrained into our daily lives, with the help of the internet and emails to complete our sentences. There is, however, a significant difference between those iterations and “generative AI” as compared to “full self-driving” such as ChatGPT. Using ChatGPT, questions can be answered in ways we once thought were the exclusive domain of humans, more quickly and at a lower price.

‍

According to a few experiments, the potential for generative AI to replace workers is astounding. According to the results of a study by Shakked Noy and Whitney Zhang, doctoral students from the Massachusetts Institute of Technology, grant writers, data analysts and human resources professionals, who used ChatGPT to produce news releases, short reports, and emails in 37% less time and with superior results, ChatGPT reduced their time-consuming processes by up to 10 minutes on average.

‍

Researchers from Microsoft Corp. recently conducted a study where they used a tool to compute the time required to program a web server using a model developed, like ChatGPT, by the startup OpenAI, and they found that the time to program the server was reduced by about half.

‍

This is a breakthrough in terms of results. A Goldman Sachs Group Inc. economist concluded generative artificial intelligence could increase labor productivity growth by nearly 1.5 percent points a year, a de facto double of what it is now by applying the same generalization to the entire economy.

‍

Automation has been dislodging labor continually for centuries, but historically it has taken its toll on routine, repetitive tasks. It means that some workers will no longer need to do a given task. By contrast, generative AI is designed specifically to target college-educated professionals right in the face of their human capital needs.

‍

The researchers at Open AI and the University of Pennsylvania asked a team of humans and a ChatGPT-like model to identify which occupations were most likely to be affected by generative artificial intelligence. According to the study, some jobs are not vulnerable to this situation. They include dishwashers, motorcycle mechanics and short-order cooks. Mathematicians, interpreters, and web designers were among the most vulnerable occupations. A study concluded that 19 percent of all workers were at risk of at least half of their jobs being affected. There is a possibility that journalists will be exposed to 100% of the risks listed above.

‍

I like to think that technological disruption is the equivalent of your neighbor losing his or her job through automation; the robot apocalypse is the equivalent of your neighbor being lost by automated technology. It is no secret to professionals, including those who write columns for their living, that blue-collar workers have been plagued with the fear of obsolescence for generations. There might be a coincidence, but it seems that layoffs these days are more prevalent among blue-collar workers than among the latter.

‍

Although some caveats need to be considered, such as ChatGPT being incorrect at times: it directed me to a nonexistent study during my research for this column. However, to ask whether ChatGPT is "right" is to overlook the fundamental point of this article.

‍

As Jim Manzi, a partner at Foundry.ai, which develops AI applications for business, explained, the aim of a large language model is not to provide correct answers, but rather to provide pleasing answers. "Its job is to anthropomorphize, to make answers that people want to hear.

‍

Whether it is a person or a robot, artificial intelligence works this way; we train algorithms based on images that are perceived as looking like dogs by humans. They aren't objectively correct or wrong in any sense.

‍

Despite the obvious polarity between ChatGPT and other predictive algorithms, the issues related to the latter are still plausible. (For example, the 1941 short story “Liar!”, written by Asimov in 1941, a telepathic robot does something that shouldn’t be done in order to avoid hurting people’s feelings.)

‍

There is nothing objectively right or wrong about a legal or medical opinion, a college course syllabus, or a column in a newspaper; so why shouldn’t a computer do the same thing? And as time goes by, ChatGPT is supposed to make fewer factual errors and to score 150 points on the SAT in comparison to its previous version. The latest version reportedly scored 150 points higher than the old one. There is even a math chatbot plug-in included with some of the versions to counteract its atrocious performance on basic algebra.

‍

It is true that artificial intelligence could be dangerous. Despite the dangers that it presents, it may mislead people, propagate divisive, hateful, and destructive speech, or even take over decisions from humans. In all cases, innovation entails some negatives, but only in few cases are these detrimental. The fact remains that, with so much money to be made and China racing ahead of the curve, these concerns are not likely to slow down development.

‍

The preponderance of evidence points in the opposite direction, so that while there are several reasons to believe that ChatGPT will wipe out more jobs than those generated by past innovations. 

‍

There are many flaws in the way predictions of how technology will affect the labor market. Experiments such as those that involved artificial intelligence can rarely be replicated in the real world. Nearly twenty years ago, some scholars estimated that one fifth of jobs in the United States would be offshored after the advent of fiber optic connections. That prediction never came to pass. There was a time when economists began warning of the potential job losses millions of high school graduates would suffer due to self-driving trucks. The demand for truck drivers has never been higher than it is today, and employers are begging for more drivers than ever before.

‍

The reason behind this is that technology is often not good enough to support the task, or the human tasks are too complex to replace them with technology. Regulation and inertia get in the way of this, so the impact unfolds over time, without being noticed among the multitude of other factors. 

‍

According to Joshua Gans, a professor of economics at the University of Toronto, who specializes in artificial intelligence, technology has transformed what was scarce into something that is abundant, and it has revealed to us what the true value of this stuff is to us as well. It is his belief that journalists will be most valuable in asking good questions and evaluating the quality of the answers instead of simply writing the results down.

‍

As a result of spreadsheets, a great deal of math-intensive analysis has become easy and cheap, and countless new occupations and jobs have resulted in these changes. In the same manner, a large language model could also lead to an explosion of applications in which large amounts of information must be synthesized into serviceable prose in order to be serviceable. Students can ask questions to the chatbot trained on course-reading materials and lecture transcripts, which will be provided to them as an adjunct to the university's teaching assistants, who have little budget for such a job. A chatbot's capability is being developed by Gans and his colleague.  

‍

In the near future, such as it did with artificial intelligence, we will see a lot of human labor becoming automated. The reason for this is that as more and more periods of human labor are taken over by machines, humans need to compete with these machines. Thus far, that has always been the case in human history.