“Like many other recent studies, we found significant potential

[tanjimajuha20]

There is a persistent belief that artificial intelligence is taking and will take jobs from people. A group of employees of the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology (MIT CSAIL), sponsored by the MIT-IBM Watson AI Lab, conducted their study of this threat. According to the researchers, there is no conflict, and people have nothing to fear.

In their study, MIT benin whatsapp number database researchers sought to estimate the likelihood of certain tasks being taken over by artificial intelligence, as well as the likelihood that businesses would incur costs to employ AI in place of humans. Contrary to widespread alarmist predictions, the MIT researchers found that most jobs that other studies have identified as being at risk of being replaced by AI are not, in fact, economically viable to automate. At least not yet.

for AI to automate tasks,” said Neil Thompson, a research scientist at MIT CSAIL and a co-author of the study. “But we can show that many of these tasks are not yet attractive for automation.” In other words, the rumors of AI replacing humans are greatly exaggerated. And if they are happening, they are not happening quickly and are far from widespread.

During the study, the researchers interviewed workers to identify tasks that AI could perform to completely replace them in the workplace. They then calculated the cost of creating an AI system to perform the identified tasks from implementation to subsequent support. In this case, only job responsibilities for visual inspection of product quality or performance of work in the field of manufacturing, construction, and everything that is not related to agriculture were studied.

The scientists used a bakery as an example to explain how the analysis was conducted. On average, a baker in the US spends 6% of their working time checking product quality. This is a task that can be automated. With the help of automation, a bakery with five employees and a salary of $48,000 could save $14,000 a year. At the same time, automation of control will cost $165,000 at the installation stage and $122,840 per year for support in all subsequent years of operation. Or even more expensive.

“Humans are still the best economic choice for these types of jobs,” the study’s authors say. Only 23% of jobs for which employees are paid could be economically automated. Moreover, even if an AI system could be found that cost just $1,000, there would be many jobs where it would be unprofitable to implement.

"Even if we consider the impact of computer vision only in terms of visual inspection tasks, we find that the rate of job loss [from AI implementation] is lower than what is already observed in the economy," the scientists continue. "Even with a rapid decline in the cost [of AI implementation] by 20% per year, it will take decades for computer vision tasks to become economically viable for companies."

The study runs counter to other predictions. For example, Goldman Sachs estimates that AI could automate 25% of the entire labor force in the next few years, while McKinsey analysts believe that by 2055, nearly half of all work will rely on artificial intelligence. Research from the University of Pennsylvania, New York, and Princeton suggests that ChatGPT alone could impact about 80% of jobs (the MIT researchers did not address ChatGPT). And an analysis of U.S. jobs suggests that AI is replacing thousands of workers right now.