Using Generative AI in Computer Science Internships; Implications for Career Readiness

Abstract

This empirical study delves into how computer science interns use generative AI tools during their internship experience and shows how these technologies connect academic training with practical professional practice. Grounded on Social Cognitive Career Theory (SCCT), we examine how GenAI usage influences students' self-efficacy and internship outcomes during internships in fast-changing tech environments. Data were gathered via mixed-methods surveys from 40 undergraduate computer science students at a large Southeastern U.S. public university, with internship experience and also worked as teaching assistants. Findings show that while GenAI got the green light in 65.8% of internships, its use was cautious and infrequent (rarely used in 38.7% of cases) and often limited to supplementary tasks such as debugging, concept clarification, and documentation summarization. A significant preference (46.7%) appeared for the company’s internal tools over public large language models like ChatGPT, due to the company’s security concerns and various company rules ranging from encouragement to full prohibition. Furthermore, qualitative insights reveal how these technologies enhanced users' productivity in supportive environments while presenting barriers in restrictive settings, which could impact students' future tech-driven careers. These results offer valuable tips and guidance for computer science educators and advocating curriculum enhancements in GenAI literacy to better prepare students for AI-integrated jobs. By addressing policy mismatches between academia and industry, we can help spread skills fairly while keeping more people engaged in computing careers.

Ähnliche Arbeiten