Sometime in the summer of 2025, Dawn DeCosta, director of partnerships at a tech-focused career-prep program at the City University of New York, started hearing employers say that they wanted computer-science students to be “AI-ready.”
“I said, ‘Yeah, do you know any AI experts? Because we’d love to have them teach,’” De Costa recalled. “‘We all need to be willing to either teach each other, or you need to give us a minute to catch up.’”
Now, over a year later, DeCosta says her program has begun to catch up.
DeCosta’s program, which began in January 2025 and operates under the city’s Tech Talent Pipeline initiative, is based out of New York City College of Technology, a CUNY school.
During the 18-month program, 30 students are given specialized opportunities, like interview training or site visits to companies like LinkedIn. They also receive additional technical training. Before, that amounted largely to learning additional programming languages. Now, according to program director Douglas Smith, the training also teaches students how to use generative AI to code, which cuts down on coding time. The students are also taught how to verify and evaluate code written by AI, a skill that experts say is increasingly important for software engineers to know.
But the opportunities for career development and AI-readiness are not uniform across the CUNY system.
According to a report released in August by the Center for an Urban Future, a New York City-focused think tank, CUNY offered only eight AI or machine-learning courses across its 26 campuses during the fall 2025 semester, meaning that most computer-science students couldn’t take those classes. Programs like DeCosta’s are also not available at every CUNY campus.
“We had a lot of outreach from other campuses who wanted to be part of this program,” said DeCosta. “And unfortunately, the way the grant was written, we could only take our students.”
If a CUNY student can’t get AI training, that can come at a high cost. Another report released in early September by the Center for an Urban Future found that the number of the city’s entry-level tech job postings have dropped in half since 2022.
“That raises the bar for recent grads,” said Eli Dvorkin, editorial and policy director at the center. “Employers increasingly expect practical experience, fluency with AI-enabled tools, and the ability to apply technical skills in real-world settings. And they’re expecting all of that while simultaneously no longer providing the entry-level paid experiences that are so essential for recent grads to get their foot in the door.”
Across CUNY’s computer-science departments, curriculum updates are happening, but not at a standardized pace.
At York College, Professor Thitima Srivatanakul piloted a class on AI-assisted coding during this past winter session. But since the class took place over winter break, only nine students enrolled. This winter, she’s considering a new class covering AI coding, along with teaching students how to make AI agents. But she isn’t sure she’ll have the time to pull it off.
Larger-scale changes face other hurdles.
“The process of creating new academic programs, given the approvals that are required at the state level, is incredibly onerous,” said Dvorkin. “That makes it a lot harder to meet the fast-changing demands of an economy that is not waiting for state bureaucrats to approve a new program over the span of, you know, 18 months or two years or longer.”
John Jay College Professor Fatma Najar noted that major curriculum changes need to be reviewed by the provost and other faculty committees. For now, if John Jay students want to learn how to use AI agents or improve their prompt engineering, Najar said, it would have to be as a self-directed project, assisted by a professor.
In some cases, students are learning AI skills on their own, through outside courses and events. Bi Rong Liu, a recent Brooklyn College graduate, said he learned to work with AI through self-directed learning and AI-focused hackathon events. At a 2025 hackathon at Columbia, Liu created an AI nutrition assistant designed to help visually-impaired users track their food and drink intake.
While he did take an artificial-intelligence class at Brooklyn College, Liu said the course was more focused on theory than practical applications.
“I think a lot of the curriculum needs to be updated, because it’s lagging behind what the industry standards are,” said Karina Lam, a recent Brooklyn College graduate.
Lam now works in IT and software engineering at CUNY. While she was a student, she tutored other students in introductory computer-science classes. When it came to AI in the classroom, beginning around 2024, she said that professors recommended it mostly as a teaching aid that could explain concepts.
She learned how to create AI agents through CUNY Tech Prep, a career preparation program. Lam said that the instructors brought in an IBM engineer to teach them about AI agents.
At present, programs like CUNY Tech Prep or DeCosta’s Tech Talent Pipeline program are one way in which students can prepare for the workforce. “I would love to see it scaled because it works,” said DeCosta.
That scale-up may be on the horizon. In May, CUNY announced a three-year initiative called “CUNY Tech Futures” that would invest over $5 million toward AI programs and degrees across 12 campuses.
“As employers raise the bar on AI and emerging tech skills requirements, CUNY’s tech degree programs have to meet the moment,” Brendan Collin, Tech Talent Pipeline Executive Director, said in the news release about the initiative.
Dvorkin realizes that faculty still face long curriculum approval times, and other bureaucratic and financial hurdles.
“CUNY Tech Futures is certainly not going to be the answer all on its own,” he said. “But it’s a really important step.”
About the author(s)
Heather Chen is a reporter with Columbia Journalism School's Toni Stabile Center for Investigative Journalism, covering technology, sports, and politics.