AI LITERACY IN UNDERGRADUATE EDUCATION: EXPLORING STUDENT PERCEPTIONS AND FUTURE CAREER IMPACT

Abstract

The rise of artificial intelligence (AI) has transformed various sectors, including education. Large Language Models (LLMs) like ChatGPT, Copilot, and the recently launched Deepseek are reshaping learning by influencing how students learn, teachers instruct, and institutions operate. However, AI adoption raises both concerns and opportunities. On one hand, overreliance on AI may hinder critical thinking, problem-solving, and classroom collaboration while also raising ethical issues such as cognitive bias, privacy, and data security. On the other hand, AI is seen as a tool to enhance learning through self-directed study, immediate feedback, and personalized educational experiences. Understanding this dual perspective is key for evaluating AI’s role in academic and professional development. This study explores how freshman undergraduate students perceive AI, focusing not on its short-term impact on learning but rather on their awareness of its potential role in their future professional careers. Assessing AI literacy among students is crucial, as AI is expected to become a relevant component of their professional environment. The research was conducted in a first-year Chemistry course within the Environmental Sciences degree at Rey Juan Carlos University. The activity, worth 10% of the final grade, followed a three-phase structure: (1) learning to formulate effective prompts and critically analyse AI-generated responses, (2) solving an ill-structured problem designed by the professor using the assistance of AI, and (3) formulating and solving a self-created ill-structure problem with AI support. Student performance was assessed using a different rubric on each stage, with average scores of 7.4, 7.2, and 7.6 in each phase, respectively. These good scores indicate strong engagement, reinforced by an 85% completion rate (79 out of 93 students). To assess students' perceptions of AI’s impact on their professional development, a pre- and postactivity survey was conducted using a Likert scale from 0 (strongly disagree) to 5 (strongly agree). Cronbach’s alpha values of 0.74 and 0.80 for the initial and final surveys, respectively, indicated acceptable internal consistency of the instrument. The statistical analysis, which included Shapiro-Wilk tests for normality, revealed no significant overall differences between the initial and final responses, as determined by the Wilcoxon rank-sum test (p-value = 0.6849). However, when analysed by gender, significant differences emerged among female students regarding the relevance of AI in their professional futures. Specifically, the Wilcoxon rank-sum test indicated a substantial increase in their final survey responses (p-value = 0.004382), while no remarkable difference was found for men (p-value = 0.09015). This shift in perception led female students to reach confidence levels similar to those of their male counterparts. These results align with previous studies suggesting greater initial reluctance toward AI among women. However, the findings also highlight how a well-structured and guided approach fosters greater confidence in AI, particularly among female students, who initially showed more hesitation but experienced a significant positive shift in their perceptions.