Abstract:

This paper presents the findings of a collaborative research study funded by an Innovation Grant from Oregon Tech Provost’s Office, aiming to examine the risks, benefits, and opportunities of integrating chatbots and artificial intelligence (AI), particularly ChatGPT, into higher education. The multidisciplinary team, comprising professors from mathematics, physics, history, art, and electrical engineering, conducted a comprehensive investigation involving literature review, question development, student recruitment, data collection, analysis, and dissemination of findings.

The study encompassed six stages: a literature review indicating widespread acknowledgment of AI's potential impact on education, particularly stressing the need for research and establishment of best practices. Subsequent stages involved faculty-generated questions for ChatGPT evaluation, recruitment of students to answer discipline-specific questions with or without ChatGPT assistance, collection and review of student responses, statistical analysis, and dissemination of results.

Key findings suggest that while ChatGPT holds potential as an educational aid, it currently demonstrates limitations in accuracy and comprehension, often producing verbose but flawed responses. However, faculty and AI detection software exhibited comparable abilities in distinguishing ChatGPT-generated responses from student-created ones. Student perceptions highlighted varying levels of confidence in ChatGPT's accuracy and identified its weaknesses, with some expressing skepticism about its future utility in academic settings.

Overall, the study underscores the need for informed discussions on leveraging AI responsibly in education and emphasizes ongoing research to better understand its implications and refine its integration into higher education practices.

Keywords:

ChatGPT, artificial intelligence, AI, AI ethics, multidisciplinary.