The Technology section is published with the support of Favbet Tech
Researchers at the Massachusetts Institute of Technology (MIT) Computer Science and AI Laboratory tried to predict , does artificial intelligence automate people's work – and if so, which one and when?
This is not the first attempt to assess the impact of modern technology on people's work. For example, a report from the investment company Goldman Sachs stated that AI could automate up to 25% of the entire labor market in the next few years. The McKinsey Global Institute says that by 2055, nearly half of all work will be done with the help of artificial intelligence. A survey by the University of Pennsylvania, New York University and Princeton shows that ChatGPT alone could impact about 80% of jobs. And the company Challenger, Gray & Christmas, an outplacement company, says AI is already replacing thousands of workers.
However, the MIT study found that most jobs previously identified as likely to be displaced by AI are not “cost-effective” for automation—at least not at this time. The key takeaway is that the expansion of artificial intelligence may be slower—and less dramatic.
«As with most recent research, we see significant potential for artificial intelligence to automate tasks. But we can show that many of these tasks are not yet attractive for automation,” says Neil Thompson, co-author of the study.
Important: The study only looked at jobs that require visual analysis—for example, checking product quality at the end of a production line. Scientists have not examined the potential impact of text and image generation models such as ChatGPT and Midjourney on workers and the economy; they leave this for further research.
The MIT researchers also surveyed workers to understand what an AI system would need to do to completely replace them, and then modeled the cost of building such a model and predicted whether businesses (particularly “non-farm” companies in the US) would be willing to pay upfront and operating costs for such a system.
At the beginning of the study, scientists analyze the forecast following the example of a baker. According to the U.S. Bureau of Labor Statistics, a baker spends about 6% of his time checking the quality of food, a task that can (and currently) be automated by AI. A bakery with five bakers making $48,000 a year could save $14,000 if it automated food quality checks. But according to researchers, deploying a “bare” AI system from scratch will cost $165,000, and annual support will cost $122,840.
Vacancies
Journalist, author of stories about IT, business and people in MC.today MC.today
Lead/Senior Quality Assurance Engineer Xpand Ukraine
Marketing Manager (e-commerce) MEDIA MINGLE, Remote
CSR Dedicated (Night shifts) #24 WOW 24-7.io, Remote
“We found that only 23% of the wages people are paid to perform visual tasks would be economically attractive to automate with AI,” Thompson said. “Humans are still the best economic choice for performing these parts of the work.”
The study further looked at autonomous AI systems sold through vendors such as OpenAI, which only need to be fine-tuned to perform specific tasks rather than trained from scratch. But even if a system costs just $1,000, the researchers say, there are many jobs—albeit low-paying jobs that rely on multitasking—that wouldn't make business economic sense to automate.
“Even if we consider the impact of computer vision only on visual tasks, we find that the rate of job loss is lower than what is already observed in the economy,” the researchers write. “Even if costs were rapidly reduced by 20% per year, it would still take decades for computer vision tasks to become cost-effective for companies.”
The study has a number of limitations, which the researchers – to their credit – acknowledge. For example, it does not consider cases in which artificial intelligence can complement rather than replace human labor (for example, analyzing an athlete's golf swing) or create new tasks and jobs (for example, supporting artificial intelligence systems) that did not exist before. Additionally, it does not take into account all the possible cost savings that can be achieved from pre-trained models such as GPT-4.
Python course. This programming varies from basic concepts to complex concepts in the spheres of individual intelligence. After completing the course, you are guaranteed to get a job in IT. More about the course
At the same time, there are questions about the pressure on scientists from the sponsors of the MIT-IBM Watson AI Lab research. The laboratory was created for $240 million by IBM, a company that is interested in ensuring that artificial intelligence is not perceived as dangerous.
Favbet Tech is IT a company with 100% Ukrainian DNA, which creates perfect services for iGaming and Betting using advanced technologies and provides access to them. Favbet Tech develops innovative software through a complex multi-component platform that can withstand enormous loads and create a unique experience for players. The IT company is part of the FAVBET group of companies.