As A.I. Majors Grow, Universities Race to Keep Curricula Relevant
Christian Vivas has spent more than 15 years running a photography and video studio and, for the past year, working as an artificial intelligence strategist at an advertising software company in Miami. Yet despite that seemingly successful career, the 38-year-old has returned to school, enrolling at Miami Dade College to pursue a bachelor’s degree in a fast-emerging field: A.I.
“To apply any type of technology that comes out, you truly need to understand what it is,” he told Observer.
Vivas is part of a growing cohort of students betting that a formal education in A.I. will help future-proof their careers. At least 70 colleges in the U.S. now offer A.I. majors, while at least 90 offer A.I. minors, according to Northeastern University. Just five years ago, only five schools offered A.I. majors, The New York Times reported. Thirty-two schools have recently graduated classes of at least 100 students with such degrees, according to Northeastern.
For universities, the rush to train students in a technology advancing at breakneck speed presents a difficult balancing act. College leaders must update curricula to reflect the latest breakthroughs while teaching the durable concepts students will need long after today’s tools become obsolete.
“A.I. degrees can be worthwhile but really only if they are delivering transferable skills that are based around real labor market value and have a foundational piece that is helpful in terms of lifelong learning,” Maria Flynn, executive director of Jobs for the Future, a nonprofit focused on addressing workforce issues, told Observer. “You can think about it as the field really being in the early innings.”
In 2018, Carnegie Mellon University became the first U.S. school to offer an undergraduate degree in A.I. Its curriculum is rooted in mathematics and statistics, which underpin the development of neural networks and machine learning.
“New types of neural networks come out [and] they have a different structure, but the same basic idea of incremental learning based on data is fundamental there, and we teach the mathematics behind that,” Reid Simmons, a professor at the Carnegie Mellon Robotics Institute, told Observer. “The idea is that if students understand the basic fundamentals upon which all of A.I. is built, then whatever new technology comes along, they will have a much easier time coming up to date.”
Simmons said he and his colleagues did not anticipate the scale of A.I.’s expansion in recent years. Carnegie Mellon has responded by continually adding courses for the 120 undergraduates currently enrolled in the program. “You could start a course in an advanced A.I. machine learning topic, and by the time you finish the course, some of what you have taught is already old hat,” said Simmons, who teaches Concepts in Artificial Intelligence and Autonomous Agents courses.
Updating courses is hardly a novel problem for computer science departments, said Chris Callison-Burch, an A.I. professor at the University of Pennsylvania.
“In computer science, we are fairly used to technology that changes rapidly,” Callison-Burch told Observer.
A few years ago, Callison-Burch created a new version of his natural language processing course. During the most recent academic year, he again overhauled roughly 40 percent of the class to accommodate “all the breakthroughs in the large language models that have happened,” he said.
Even as universities race to keep pace, the outlook for the labor force remains unsettled. Recent graduates with degrees in computer engineering and computer science posted among the highest unemployment rates of all majors, according to the latest data from the Federal Reserve Bank of New York. The bank does not track A.I. degrees as a separate category.
“The worry among many undergraduates is that the entry-level jobs that they normally would have found at a big tech company like Google or Amazon can now be done by A.I. agents,” Callison-Burch said.
Yet Simmons said Carnegie Mellon’s results offer some reassurance. Among graduates who remain interested in an A.I. career, nearly all have found jobs, including at OpenAI, Google and ByteDance. Others have chosen different professions or continued on to advanced degrees, he said.
Callison-Burch is similarly optimistic about Penn graduates’ prospects, even if the market looks different from the one that greeted earlier generations of computer science students.
“The amount of software being created is growing at such a huge rate that my strong belief is there is going to be strong demand for people who understand A.I. and people who understand computer science,” Callison-Burch said. “But those jobs may be spread across a much wider range of employers than they were before.”
