ChatGPT will affect individuals holding bachelor’s, master’s, and professional degrees more than those without these credentials, according to a paper published last week.
A group of researchers from OpenAI and the University of Pennsylvania found that jobs with the least exposure to AI require the most on-the-job training. Jobs that don’t require much training are associated with higher incomes—but are also more exposed to GPT.
Overall, OpenAI’s study found that about 80% of the US workforce could have at least 10% of their work tasks affected by GPTs, while around 19% of workers may see at least 50% of their tasks affected.
“These models have made a lot of progress in understanding human speech, both written and spoken and using pictures,” said Simon Johnson, an economics professor at the Massachusetts Institute of Technology. “They are able to potentially replace humans in tasks that are fairly intensive in communication—and a lot of communication tasks are done by people with some degree in education.”
AI taking away jobs is only half of the story
Johnson said that while it’s possible many white-collar workers will be displaced by AI, it’s also possible that the technology will create new ones. The question is: Who can adapt, be flexible, and learn new skills successfully enough—and quickly enough—to take advantage?
Since the 1980s, Johnson said, many new Internet-related jobs were formed, such as web design and marketing, while jobs that entailed clerical tasks such as data entry diminished. Many of the new jobs cratered to younger people, those with an education, and those with a different perspective, he said.
More broadly, AI has been showing up in the workplace for years, whether it’s assisting with warehouse inventory, automating customer service in call centers, or helping companies gain insights by analyzing large swaths of data.
Another reason for why AI could affect white-collar jobs the most? Pieter den Hamer, an analyst at market research firm Gartner, told the Washington Post that AI can be applied at a relatively low cost to a white-collar job, while deploying a fleet of autonomous trucks, for instance, would be highly expensive.