Carlmont Advanced Placement Computer Science students take a test in teacher Jadie Sun’s classroom. As students work through the assessment, they apply the programming concepts and problem-solving skills they have developed throughout the course.
Artificial intelligence (AI) can now write and debug code in seconds, threatening many jobs in the computer science field. As AI tools become more common in software development, concerns about entry-level opportunities have led some students to reconsider pursuing a career in the field.
JJ Furman, a Distinguished Software Engineer at Google who helps build AI systems for Google Search, said the capabilities of AI coding tools have advanced rapidly.
“Even just a year ago, these AI programs could barely write small pieces of code,” Furman said. “Now they’re responsible for almost all the code being written.”
Despite these advances, Furman said AI-generated code is not always reliable. Developers still need to review its output to ensure that the code works as intended.
“One of the biggest problems with AI is that it makes code really fast, but it’s not always the best,” Furman said.
A 2025 Stack Overflow Developer Survey found that 46% of respondents distrusted the accuracy of AI-generated output, compared with 33% who trusted it.
Furman said relying on AI without understanding its output can create safety and security risks. As AI takes over routine coding tasks, programmers will still need to understand how programs function and how their components fit together.
“The coding itself is the part that AI automates,” Furman said. “What humans still have to do is make the big plans and piece all the little things together at the higher level, while reviewing the code.”
These changes have also raised concerns about entry-level opportunities. A discussion paper from the Institute of Labor Economics found a 14% to 15% relative decline in junior software developer vacancies compared with senior positions following ChatGPT’s release.
At Carlmont, Advanced Placement Computer Science A teacher Jadie Sun said students entering the field may need to demonstrate skills beyond writing basic code.
“To work in computer science, learning how to code is now just a basic step,” Sun said. “AI is doing a really good job of writing advanced code, but humans still need to review the code.”
Sun said students may need to distinguish themselves through self-motivation, practical experiences, and participation in activities such as hackathons to prepare for internships and entry-level positions.
The growing capabilities of AI have also affected how Sun approaches teaching.
“I expect that anything that’s not done in class might not be their own work. So I make sure to give my students the opportunity and time to complete their work in class,” Sun said.
Sun said she doesn’t want her students to have to rely on AI systems to complete their work, while also acknowledging the pressure that easy access to AI places on students not to use AI to code for them.
“I give my students opportunities to come to me with questions during class, so they don’t feel the need to use AI for their assignments,” Sun said.
For students considering computer science careers, the uncertainty extends beyond the classroom.
Carlmont junior and Computer Science Honors student Ellen Povzner is interested in pursuing computer science but worries about the field’s long-term sustainability and the competition for jobs.
“Since AI is taking a bunch of jobs in computer science, it could be difficult for me to excel, given the high competition for top roles,” Povzner said.
Despite these concerns, Povzner said she has used AI as a tool to support her learning, acknowledging its benefits as a teacher and assistant.
“Even with the rise of AI, I still think it’s very important to learn these skills, especially to understand what’s actually happening behind the AI,” Povzner said. “Personally, I’ve used AI to help me learn to code in class and to give me tips and tricks. I find that it can be a helpful learning tool.”
Furman similarly encouraged students to use AI to strengthen their understanding rather than simply automating their work.
“Doing the tasks is the way you actually learn. I wish students and programmers would use AI as a teaching aid rather than a labor-saving aid,” Furman said.
Povzner remains interested in exploring computer science classes and potential careers, but the uncertainty has also led her to consider other fields.
“I’m still excited to explore classes and potential jobs in computer science since I enjoy coding and programming. But I am a little concerned, which has led me to explore different careers,” Povzner said.
As AI continues to evolve, students like Povzner must weigh their interest in computer science against uncertainty about the field’s future.
“We don’t have all the answers to questions about AI and its future right now,” Furman said. “AI is ever-evolving, and who knows what it could do in the future?”