Are any jobs safe from AI?

SALT LAKE CITY — These days, a lot of workers have been asking themselves whether AI is coming for their job.

Computational social scientist Isabella Loaiza saw that fear in the people around her and with another MIT researcher, economist Roberto Rigobon, looked for answers – not from AI, but from the human side of the equation.

Working with a U.S. Department of Labor database, O*Net — a list of more than 900 jobs and 19,000 individual tasks performed at those jobs — they devised what they called an EPOCH score, a measure of human traits needed in each job – E for empathy, P for physical presence, O for opinion or judgement, C for creativity and H for hope or leadership.

They came up with a measure of how much AI might be used to augment each job and then calculated a measure of the risk of being replaced by AI.

Jobs in caring — nursing, social work, home care, health care practitioners — had a lower risk as did jobs that require a lot of cognitive flexibility like emergency management directors, who need to make difficult decisions in a short amount of time. Jobs that call for social skills — like kindergarten teachers and CEOs — also did well.

At the other end of the scale were jobs with a lot of paperwork — mail clerks, meter readers and tax preparers.

Surprisingly, some real hands-on jobs like flooring installer and plumber didn't do well, but Loaiza said that's because the O*Net database doesn't always include the obvious information that those jobs require the person to be on-site.

Loaiza said because those jobs are done in-person and can't yet be replaced by machines, they're safe for now.

In the research paper The EPOCH of AI: Human-Machine Complementarities at Work, researchers said AI is more likely to augment jobs rather than replace them.

But are there any AI-proof jobs? Loaiza said that's a question she can't answer.

"I think it's hard to predict the future," she said.

She did say that the occupations with higher EPOCH scores are jobs that, according to information from the Bureau of Labor Statistics, have done well and continue to thrive in the workplace.

So what's the takeaway?

Loaiza, educators and professionals we spoke with offer two pieces of advice.

1. Embrace (or at least learn) AI

Aimee Watson was studying art at the University of Utah until she saw images produced with Midjourney AI.

"Most of them were visually above my skill level and only took a matter of minutes to produce versus the hours to days I would have to spend making something comparable," she wrote. "Career-wise, I wasn't sure how I would be able to compete."

She switched to Film and Media studies which, especially with the release of new AI tools and services, is now also threatened by the new technology.

Matt Hoffman, an artist at BluFire Studios, a graphic effects and CGI company in Orem, has spent countless hours creating dragons and other creatures and visual effects for the movies.

Now Open AI's Sora 2 can conjure a dragon in just a minute or two (albeit one that's not quite ready for the big screen).

"The problem, too, is that because it's trained on all of the very best $30 million dragons, it looks like those dragons. You know, it has all of the details and all of the things that typically over the course of decades, have cost hundreds of millions of dollars to produce," Hoffman said. "We saw a film (with) a building collapse, for example, and those are very, very complicated sequences. With traditional tools, it takes an enormous amount of time and money to do, and they were able to do it with generative AI very quickly. Netflix, I think, said, 'Hey, just so you know, we used AI to make this building collapse sequence.' And they said it cost three times less."

"The despair that I've seen in, in all of my artist friends over the past two years has been incredible and awful," he said.

Aimee Watson is now an assistant professor in Kenneth Collins' AI filmmaking class, perhaps, Collins said, the first of its kind in U.S. higher education.

Students learn to use AI to write scripts, create characters, video, audio and music.

"I think a student in film and media arts needs to at least know what's coming, even if they decide not to use it," Collins said.

Computer science student Ishan Sharma has shifted his coursework away from standard programming.

"A lot of my focus to taking classes that are more in the AI field — so machine learning courses, courses to do with large language models, definitely," he said.

A recent study out of Stanford University, Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, found that IT workers 22-25 in the jobs most vulnerable to being replaced by AI have experienced a 13% relative decline. It has been cited as evidence that AI has already started taking jobs.

But Mary Hall at the Kahlert School of Computing isn't convinced that is happening just yet.

"It is not clear to me that that's why, because there are a lot of economic forces that were happening," she said. "I think it's very disruptive. And so some people may have to develop new, but it's like disruptions we've had throughout the history of computing."

"AI can generate a lot of code, but whether it's correct is going to take a human to verify," she said.

Hall recommends computer science students apply for internships, take classes in machine learning and consider pursuing jobs in financial services, banking and engineering.

She said computer science students looking for jobs in those areas are doing well.

"(Those companies) kind of swept, swooped in, and they were really happy that they finally could get these employees and they weren't competing with big tech," she said. "Think about the things besides computing that you're passionate about and get some training in that, like a minor, a certificate, or maybe even a master's degree in a different area."

University of Utah law students now learn to use AI tools like Nexis +.

"It's like that movie Hidden Figures about NASA," said Bennett Borden, attorney and data scientist at Clarion AI. "It used to be that humans did all the calculations by hand. And so what do they do in this particular movie? They learned Fortran, right? That is what younger people are going to have to do to differentiate themselves in a market. If you can bring that skill set with you, you will just set yourself apart."

2. Figure out what AI can't do

English professor Hollis Robbins, who writes about AI in higher education in her Substack newsletter, Anecdotal Value, said students should figure out what AI knows and what it doesn't know.

"I met a young man recently who wants to go into sports journalism," she said. "Well, all the facts, the statistics … is … on AI. But what do you know about the game? What do you know about why people go? What creates excitement? That's your job. That'll be your career right there. If you can situate yourself and your brain to understand, there are things that I bring to the table that this machine cannot, then I have a place in this world."

Learn AI and invest in the things that make you human, the components of that EPOCH score, Loaiza said.

"You probably won't be replaced by machines just yet, but you will be replaced by another human who can work with machines," she said. "If you want to make sure that you are more competitive in the future labor market, you want to invest in these EPOCH capabilities and also in your ability to work with AI."

Related topics