Underemployment Isn't New

Last year a study by the Burning Glass Institute and Strada came out called “Talent Disrupted”. In the study they discuss a concept called ‘underemployment’, which is when college graduates work jobs that do not require college degrees such as retail, food service, admin support, etc. There are two kinds of underemployment, severely underemployed, which is when college graduates are working jobs that require no more than a high school diploma, and moderately underemployed, which are jobs that require more than a high school diploma but less than a bachelor’s degree. 

Unsurprisingly, the degree you pursue in college is a huge determining factor in the likelihood of being underemployed post-graduation. Degrees like nursing, engineering, and math-intensive business degrees like finance and accounting boast the lowest percentage of underemployed graduates whereas degrees related to public safety, recreation and wellness, and business fields like marketing and HR have some of the highest rates of underemployment. The degrees that involve a lot of math and quantitative rigor are the ones that see the most success in college-level employment after graduation. 

But will AI change this? I believe so. I think that with the integration of AI into the corporate world, degrees with relatively low rates of underemployment will see an increase in underemployment rates. Computer Science degrees are what first come to mind. While not perfect, ChatGPT can churn out lines of code much faster than a human can. I’ve even seen someone use ChatGPT to code the game ‘Snake’. It was an early version of ChatGPT, and the game didn’t exactly work well, but it was put together much faster than a human could have. Some of the degrees and skills that were touted as ‘safe’ in the study are the very things AI does well.  

The study also puts forth suggestions for students, schools, and policy makers to help avoid underemployment, and the biggest one they have is offering access to paid internships for every college student. However, internships are already getting harder and harder to find as AI is being used to do the grunt work that interns typically would do. Other recommendations in the article do hold up better. The study recommends that universities should receive funding to give students clear and personalized guidance about finding college level employment after they graduate and what to do to avoid underemployment, including engaging in campus leadership experiences, GPA, and choice of majors and minors. This, to an extent, goes hand in hand with the study’s final recommendation of policy makers removing disincentives for public colleges to offer degrees that have low rates of underemployment. Many of the majors with low underemployment rates, like data science, engineering, and nursing, cost more money to offer than degrees with high rates of underemployment. Investing more money in public colleges and universities will give more students access to the majors that will give them college level employment after they graduate. 

Despite this study only being a year old, the rapid advancement of AI has already negated much of what the study was saying. Computer science majors are already having a harder time finding jobs as generative AI takes over the responsibilities of junior coders. 

However, I believe that, in the near future, degrees that are more human focused, like psychology and public safety and security, will become more valuable. Degrees that put more of an emphasis on critical thinking will be viewed as more valuable than the more quantitative focused degrees. Certain fields with low rates of underemployment, like nursing, and education, will stay safe as AI expands. The fields where AI can be used as a tool rather than as a replacement will continue to flourish and thrive as the technology expands.