The majority of US businesses cannot afford to replace human workers with AI computer vision due to the high expenses of the technology now in use. But in the long run, it might alter.

According to certain Fortune 500 CEOs and Silicon Valley executives, artificial intelligence may soon replace many workers; nevertheless, the cost of AI computer vision technology is now too high for the majority of US enterprises. The conclusion is based on research on human labor that is susceptible to machine automation, particularly visual tasks.

Researchers examined whether it makes financial sense to use current AI computer vision to replace vision skills found in a variety of human employment. "There are several

It was shown that while 36% of US non-agricultural enterprises had at least one worker duty that might be automated with AI computer vision, only 8% have a task that can be automated with AI at a reasonable cost.


Additionally, they came to the conclusion that companies would not find it cost-effective to automate more than 0.4% of the salary and benefits paid to non-agricultural workers in the US.

With the present prices of AI computer vision, fewer than a tenth of the existing vision functions might be cost-effectively automated, even by huge US enterprises with 5000 or more people (larger than 99.9% of all US companies).

According to Queen Mary University of London's Gino Gancia, the results showing that AI computer vision is too expensive for the majority of US enterprises "might sound like a reassuring result" but "there might be other [AI] applications with lower automation costs."

The amount of jobs available and the pay of human freelancers on websites like Upwork have already been badly impacted by the drive to adopt "generative AI," which can produce new material. Additionally, Gancia's research has demonstrated that higher employment losses have already occurred in US states like California that have sectors embracing AI more swiftly.

"We know that new technologies diffuse unevenly, generally speaking," adds Gancia. Therefore, it is likely that AI and automation would increase inequality between businesses and workers.

In the long run, Thompson and his colleagues do anticipate a sizable portion of human labor being mechanized. However, that relies on how fast the price of building and training AI systems can come down.

Governments need to begin preparing for the significant degree of automation that is expected, according to Thompson. "However, we have enough time to start implementing effective programs that will help [displaced] workers."