Systems that use artificial intelligence can write many things far faster than people can—and that includes computer code.
So what does that mean for the education of tomorrow’s computer scientists?
At Fordham, faculty made one thing clear: AI can’t beat HI, or human intelligence, the kind of deep knowledge and judgment that the computer science curriculum develops. In research labs and in classrooms, students are learning how to harness the possibilities of AI—while also avoiding its failings and pitfalls.
Understanding how AI or any other program falls short is “one of the things that we teach, foundationally, in computer science training,” said Daniel Leeds, PhD, chair of the Department of Computer and Information Science.
Using AI Responsibly
Across the department’s programs, “we have a variety of instructors who are guiding students in the proper use of AI for responsible programming,” Leeds said. That means conveying awareness of its faults—like how it can produce computer code that “looks rather good … but is wrong,” he said.
“If you understand programming, you say ‘oh, I see what it was trying to do’” and identify the mistakes, he said.
At the same time, students are being prepared to build the better AI systems of tomorrow. Students are doing that by getting grounded in fundamentals like data mining and machine learning, like they always have—“AI is built on top of computing systems, and so the standard understanding of computing systems will still be important for computer science training,” Leeds said.
The introductory courses in artificial intelligence, both undergraduate and graduate, have drawn strong interest from students for years—since long before the recent surge of public interest in the topic, said one of the course’s instructors, computer science professor Damian Lyons, Ph.D. He called AI a “dynamic and fast-moving” field that has evolved a lot over the past decade.
Recent advances in AI have spurred new graduate courses and programs at Fordham, such as an advanced certificate in cybersecurity and AI. Artificial intelligence is also a major research interest among faculty, with clusters of experts focusing on how it relates to cybersecurity and other areas, said computer science professor Gary Weiss, PhD. “Our department as a whole has a very large focus on AI,” he said.
AI in the Research Lab
Students are also learning about AI by working with faculty on research projects, which sometimes expose more of its flaws—and its biases. For instance, one student showed how AI-driven image generation programs are biased against heavier people; other projects are focused on AI’s biases based on race, gender, and disability, Weiss said.
Another project gave students a look at an innovative use for AI—combining it with multiple machine learning tools as well as traditional math disciplines like geometry to produce results more efficiently.
It’s a method that resembles the human mind’s way of combining different modes of thinking to solve a problem, said Frank Hsu, PhD, Clavius Distinguished Professor of Science. “AI without … human intelligence and natural intelligence is very shallow,” he said.
The method was developed in the department’s Laboratory for Informatics and Data Mining, said Hsu, head of the lab.
A senior working as an assistant in the lab, Yiwei Sun, noted how this method can harness AI systems for better decision making in drug discovery and other fields. But in the lab, she has also learned about AI’s faults—ironically, as when an AI query tool gave the wrong inventor for the method, combinatorial fusion analysis, which Hsu developed.
She also has seen how AI can “hallucinate” when used for research. But, she said, “I can use AI as a tool that helped me understand some simple concepts” in programming.
Students, both undergraduate and graduate, will also learn about combining AI with quantum computing through a new seminar co-organized by Hsu and Samuel Chen, MD, PhD, a scientist at Wells Fargo who spoke on this topic when he recently delivered the Clavius Distinguished Lecture at the Rose Hill campus.
An Evolving Technology
Referring to reports of disruptions in the hiring market for computer science majors, Weiss noted that computer science has always been cyclical, with surging interest in Silicon Valley followed by AI “winters,” or slumps.
The logic-based AI systems he learned about as a student are a far cry from today’s generative systems that read text, spot patterns, generate answers to questions—and sometimes get things wrong, he said.
Messing up sentences is one thing, he said, but “it could be a bigger deal in [computer]code, because it could cause a big problem.”
