Janice R. Andrews
One of the most enduring worries in the era of quickly developing generative AI is whether or not this technology would replace human occupations, particularly in the field of software coding. With programs like GitHub Copilot that can write large quantities of code, many people question whether getting a computer science degree is still beneficial. A convincing response to this topic is provided by new data from the University of California, Berkeley, which indicates that students are still drawn to computer science.
An Increase in Attention*
Students’ interest in computer science is still growing in spite of the developments in generative AI. The number of first-year applications to the College of Computing, Data Science, and Society (CDSS) at UC Berkeley, one of the best universities in computer science, has increased significantly. For the Fall 2024 incoming class, applications for these majors increased from 9,649 to 14,302 (a 48% increase) over the previous year. On the other hand, UC Berkeley’s total first-year applicant volume was largely unchanged.
Persistent Passion for Careers in Computing*
Professor Jennifer Chayes, dean of Berkeley’s College of CDSS, brought attention to this increase in applications at a recent fireside discussion with Stanford Professor Fei-Fei Li and Governor Gavin Newsom at the Joint California Summit on Generative AI in San Francisco. This research shows that students are still motivated to pursue professions in computer science, even in the face of worries about how AI may affect the labor market.
Human Developers’ Role
I got in touch with John DeNero, chief scientist of the generative AI firm Lilt and a computer science teaching professor at UC Berkeley, to learn more about this trend. Although students voice some concerns about how AI may disrupt the software engineering job market, DeNero—who has previously worked on Google Translate—emphasized that students are still enthusiastic about jobs in computers.
“Many of the challenging aspects of software development can’t be reliably performed by generative AI at this point,” DeNero said. “There will still be a central role for human software developers long into the future.”
Generative AI’s Limitations
While generative AI is great at mimicking existing software, it is not as good at coming up with completely original and creative solutions. This is especially valid for creative software development projects requiring a high level of creativity. Artificial intelligence (AI) models are trained using pre-existing data, and they frequently perform poorly if the required data is missing from their training dataset.
“Generative AI requires a lot of thoughtful human intervention to produce something new,” DeNero stated. “All consequential software development projects involve quite a bit of novelty, which currently requires clever and well-trained people.”
Efficiency Instruments, Not Substitutes
DeNero also made the point that although generative AI might speed up the less exciting parts of software development, human knowledge is still required. Rather, it functions as a potent efficiency instrument that human developers might utilize to augment their output.
Practical Uses: Lilt’s Experience
At Lilt, which is developing an AI platform for translators, this dynamic is clearly visible. Even with the accessibility of programs like Google Translate, linguists are nonetheless necessary to produce translations that are of a high caliber, particularly for significant publications. Although Lilt’s task-specific generative AI models facilitate faster and more accurate work for human specialists, human judgment still plays a major role in the final outcome.
“To reliably produce publication-quality translations, human expert linguists are still at the center of the process,” DeNero stated. “Using Lilt’s AI models, those experts are much faster, more accurate, and more consistent, allowing more text to be translated at higher quality into more languages.”
Software Development’s Future
According to DeNero, there will be a similar trend in software creation, where a tiny group of really talented human engineers will be able to produce even better software with the aid of AI technologies. This implies that there will be a constant need for qualified computer science graduates, giving aspiring professionals plenty of chances to put their knowledge to use and advance the area.
The increase in applicants to computer science degrees at UC Berkeley shows that interest in computer science is still strong even with the emergence of generative AI. In order to provide novel solutions, human developers are still needed, and artificial intelligence (AI) is a useful tool that may complement rather than completely replace them. Aspiring computer scientists still have a lot of chances in front of them to use their skills to progress technology.
