|
Getting your Trinity Audio player ready...
|
The number of undergraduate computer science bachelor degrees conferred by Stanford University in 2026 decreased by 70 people – almost 14 percent – compared to the year before.
A one-year drop at a single university is not a trend. There has always been “natural variation” in the number of degrees conferred in the program, said Stanford computer science chair Mehran Sahami.
“We’ve seen these kinds of cycles before,” Sahami said. “In the long term, though, I do think that people with computing skills will still be in high demand.”
Still, computer science degrees and learning to code have been viewed as a sure path to job stability for decades. One recent academic paper from researchers at the Census Bureau found that college graduates in highly “AI-exposed” fields like computer science saw labor market outcomes deteriorate more after the introduction of ChatGPT than people in fields less-exposed to AI. AI-exposed graduates entering the job market after ChatGPT are less likely to find employment and more likely to begin careers in lower-paying careers, the paper found.
“These are quite large changes,” said Lee Tucker, co-author of the paper and a Census Bureau economist. “We think of these as being on the order of what previous literature has suggested would occur in a large recession.”
Nationwide, enrollment in computer science has taken a hit in recent years. The number of students enrolled in a computer science major dropped by 8 percent, according to an analysis from The Washington Post.
Two primary factors have driven the decrease in computer science enrollment nationwide, Sahami said. Widespread layoffs from large tech companies like Meta, Amazon and Oracle have meant that recent computer science graduates are competing with mid-career employees for jobs while hiring has slowed. Additionally, AI is being used to write more code.
Computer science graduates are stuck with the boom-and-bust cycles of technology companies. During the dot com bubble in the early 2000s, for example, enrollments in computer science dropped notably a few years later, Sahami said. That was due to a perceived decline in the financial prospects of the discipline and to other factors like labor offshoring. (The 1990s dip in degrees is more difficult to understand, but Stanford computer science emeritus Eric Roberts suggested that departments reached teaching capacity due to high demand for computing skills, and departments had to restrict admission to the field).
Despite the bumpier job market for new grads, it’s not all doom and gloom for computer scientists. Computer science and other STEM degrees remain among the most lucrative fields. Additionally, Sahami said that downturns in tech have historically been a good time to enter the field.
“If you imagine someone who was going into college in the year 2001, 2002 – when computer science enrollments were dropping – that actually would have been a great time to major in computer science because you would have come out around 2006, 2007,” Sahami said. “That’s the time when companies like Facebook, now Meta, were being founded. It was still the early days at companies like Google that had huge growth trajectories in front of them.”
By 2006, a software engineer was the top-rated job in America by Money magazine.
Lawrence Warren, another economist who co-authored the Census Bureau paper, said it’s too early to tell what the long-run impacts of AI will be. Large Language Models and other AI technology are constantly changing, for example, and economists don’t yet know how long the AI disruption will occur.
“There are a lot of ways that AI might shape the labor market in the long run,” Warren said. “We may see productivity gains and income effects…but it’s still too early to really know.”
Sahami believes that it will take time for companies to absorb the “AI dividend” – the increase in productivity that AI adds for software engineers and computer scientists – but that this productivity increase will eventually be baked into companies. Sahami’s son is even studying computer science in college.
“There is always going to be a need for good software engineers who are going to know how to use the AI tools better and be more productive,” he said.
