Recently, artificial intelligence (AI) technologies, including generative neural networks, have become increasingly widespread among the general public. One of the most advanced and widely known examples is ChatGPT, built on the GPT-3.5 architecture, which allows it to process and analyze natural language and provide accurate and often unexpected answers to a variety of questions. ChatGPT can be used in many areas, including education, healthcare, banking, and many others.
Despite all its capabilities, and contrary to numerous comments on the internet, ChatGPT, like any other “smart program,” cannot completely replace humans in the workplace. In this article, we will discuss why human labor will remain relevant.
The cost of innovation. The introduction of technologies that partially replace people in business processes is costly. In times of crisis, innovation costs are reduced, and some do not provide for them at all, so the need for human labor only increases at such times.
Economic justification. Once innovations have been created and there is a clear understanding of how to implement them, dry figures come into play. Capitalists, pursuing the goal of increasing profit margins, replace human labor with machines only when the machines ultimately require less expenditure. Economic justification is calculated for a specific task in a specific type of business activity.
Human labor cannot be completely replaced by definition. In the production process, new value is created precisely by human labor, since it is capable of not only creating its own value (labor power) at a given moment in time, but also creating surplus value, which is transferred to the capitalist. Let’s dwell on this point in more detail.
How it works in an ideal world
While the first two reasons are clear—the cost of innovation and economic justification can be calculated—the third reason is somewhat more complicated. No machine can operate completely autonomously. Suppose we replaced the only copywriter at an imaginary marketing agency with ChatGPT. What would we get? To make ChatGPT work, it needs to be given specific instructions in natural language, known as a “prompt.” Next, the text or image created by the generative model must be checked for suitability for use. In the case of text, all facts and statements must be checked for accuracy. OpenAI, by the way, does not hide the fact that “ChatGPT may provide inaccurate information about people, places, or facts.” Naturally, the above tasks still require human intervention. Let’s assume that our imaginary marketing agency had not one but three copywriters. Then, by implementing ChatGPT, it might be possible to reduce the number of copywriters by one, or perhaps even two. However, the remaining copywriter No. 3 would have to work for three, even if they did less of the “dirty” work of writing text, but the tasks of post-editing and proofreading would increase significantly.
Every innovation that is implemented requires support. If you have implemented a conditional “1C” at your enterprise, you need to at least configure it, pay for a subscription, and, if necessary, pay a consultant to help you figure out how to work with the system. The same applies to ChatGPT and other AI products. Undoubtedly, more and more jobs will be created in the field of AI consulting.
What is happening in reality
A significant number of studies have established that there is a correlation between the number of innovations and unemployment. The more business processes are “optimized,” the fewer jobs there are. Based on these findings, the hype surrounding ChatGPT has given rise to many opinions about replacing people with AI and calls to halt research.
To counter this trend, an information campaign called “AI will not replace you. A person using AI will.” Its essence is to motivate people to use and adapt AI in their daily work, since those who know how to use such things will be more productive and attractive to employers. Obviously, this campaign is pleasing to developers and consultants in the field of AI.
There is another opinion that the introduction of AI, like other innovations, on the one hand, provokes a reduction in the number of jobs requiring low and medium qualifications, and on the other hand, an increase in the number of jobs requiring high qualifications. To be more precise, it is not so much qualifications that are growing, but rather skill levels: roughly speaking, to create a website, you don’t need to know how to write machine code, you just need to know how to use an online builder.
Conclusion
Over the past 100 years, there have been several changes in technological structures and, consequently, several waves of innovation. However, there is no trend toward an increase in the average unemployment rate (at least, if we look at the statistics).