Jobs, Wages of College Graduate Workers Will Change Most Due to AI Like ChatGPT

• Since the 1980s, automation has replaced middle-income workers with robots and software.
• Automation has also increased the value of educated workers and contributed to income inequality.
• Artificial intelligence can make educated labor cheaper and reduce inequality.

Automation has increased income inequality in the United States. The latest developments in AI may turn that around.

In a study published in March, OpenAI, the company that created ChatGPT, found that modern artificial intelligence could have a greater impact on white-collar workers than the rest of the workforce. By shifting more jobs to middle-skilled workers, this technology could curb the explosive wage growth experienced at the expense of their blue-collar counterparts and reduce income inequality. But that rosy future depends on how AI develops and whether we regulate it.

So far, automation has mostly hurt blue-collar workers. MIT economist Daron Acemoglu found that automation has caused a 50% to 70% increase in US wage inequality since the 1980s. Machines, robots, computers He software has replaced workers without a bachelor’s degree in routine tasks such as assembly line work and administrative office work. Those workers with low demand have seen real wages decline or stagnate, especially those who previously earned middle-class incomes.

This automation places an emphasis on skilled white-collar workers and widens wage inequality. MIT economist David Ortow said that while automation devalued and replaced low-skilled jobs, workers with bachelor’s degrees (a substitute for high-skills) and workers with high school diplomas found that the income premium between

“Not only does it replace you, it makes your expertise more valuable,” Autor told Insider.

Replacing low-skilled human labor with machines will increase demand and wages for abstract tasks that cannot be easily automated. As a result, wage inequality has widened significantly over the past half century.

Today, the same skilled white-collar workers who were the beneficiaries of inequality have been placed in the crosshairs of artificial intelligence.

“The general speculation is that AI, and office software before it, will also impact middle-skilled workers,” Acemoglu told Insider.

“We are really in another era,” Autor said.

Previous software relied on scripted instructions to perform routine tasks. For example, recipes that result in the same meal. You can train AI to come up with answers to data you’ve never seen before. Bicycle, he said.

He added that it is “powerful and opaque” because now “computers know more than we do.”

According to OpenAI’s March paper, a large-scale language model (the type of artificial intelligence behind ChatGPT that can process and generate text that sounds very much like natural human language) has a capacity of at least 80 We found that it affects 10% of tasks for 10% of labor and 50% of tasks. At 19% of the workforce, “high-income jobs may increase exposure to LLM capabilities and LLM-powered software.” We also found that people holding a were more likely to be affected.

The term “exposure” means that AI can either reduce inequality by replacing high-income workers, or increase productivity that further increases inequality.

If the technology is as capable as engineers expect, AI could “reduce inequality” and “be a force of equality,” replacing highly skilled workers. , can reduce the need to pay them higher wages, Acemoglu said.

Even if technology is not advanced enough to replace these high-skilled workers, it can improve the productivity and quality of middle-skilled jobs, thereby boosting middle incomes.

“It’s not hard to imagine delegating some of the most highly skilled tasks to less skilled people,” says Autor. For example, with AI, nurses are “enhanced by this tool so they can do things that doctors traditionally do,” he added.

AI reducing inequality “isn’t a ridiculous scenario, but there are many uncertainties,” said Acemoglu.

“We don’t yet know the full capabilities of large language models,” he added.

“This may turn out to be a technology that widens inequality rather than harming more skilled and well-paid workers,” he said. He gave two reasons. I was.

First, “we tend to underestimate how complex human work is,” he said.

Additionally, “more skilled workers may be able to protect themselves,” he said. .

“They won’t be hit as hard as the blue-collar line workers when the robots were introduced,” he added.

“It’s a new game, and I don’t know what the new game will be,” said Acemoglu.

Autor believes AI “could make a plausible contribution to reducing inequality” among those displaced from middle-income jobs, including how workers are trained, how employers He said there would need to be changes in the way people are hired and how governments guide innovation. It doesn’t just automate things. ”

At the moment, “incentives are not well aligned with social objectives,” he said.

“This is all being pursued in an arms race by many commercial companies,” he added. “There is no reason to think that what is good for Amazon or Google is good for humanity.”

Nonetheless, he said, “the future is a choice,” influenced by how the tools are deployed.