ChatGPT Research

Overview:

  • Title: Educators’ Opinions on ChatGPT Use for Computer Science Education
  • Research Advisor: Parastoo Abtahi and Andrés Monroy-Hernández, Princeton University, Department of Computer Science
  • Timeline: Aug 2023 – Dec 2023

Abstract:

This research investigates Computer Science educators’ perspectives on students’ use of ChatGPT in completing coursework. Understanding educators’ sentiments allows us to develop stronger AI tools for CS education as well as isolate concerns as directions for future research. To accomplish this, we conducted semi-structured interviews with 7 educators. We then used thematic analysis to reveal key themes in their opinions on the use of chatbots in the classroom. We present four key themes: context sensitivity, opportunities for learning, classroom adaptations, and concerns about long-term impact. Our results call for the continued development of AI-informed classroom materials and task-specific AI tools to responsibly integrate ChatGPT in education.

The Problem

  • Understanding Educators’ Sentiments: Gaining a comprehensive understanding of the diverse views among educators on the use of AI tools like ChatGPT in coursework.
  • Identifying Opportunities & Concerns: Recognizing the balance between potential benefits and challenges presented by AI in education.
  • Developing Actionable Insights: Translating interview data into practical recommendations for integrating ChatGPT into education responsibly.

Research Questions:

  • RQ1: What are CS educators’ positive thoughts on students using ChatGPT for completing coursework?
  • RQ2: What are CS educators’ concerns about students using ChatGPT for completing coursework?
  • RQ3: In CS educators’ opinions, in what scenarios could CS students benefit from using ChatGPT?
  • RQ4: In CS educators’ opinions, in what scenarios could CS students encounter obstacles because of ChatGPT use?

The Methodology:

  • Semi-Structured Interviews: Conducted interviews with seven educators from the CS department to gather qualitative data on their perspectives.
  • Thematic Analysis: Analyzed responses using thematic analysis to identify four key themes: context sensitivity, opportunities for learning, classroom adaptations, and concerns about long-term impact.
  • Collaborative Analysis & Synthesis: Worked closely with peers to discuss and synthesize results, leading to actionable insights on AI integration in CS education.

The Findings:

  • Context Matters: ChatGPT’s use depends on task type, performance, and learning goals. It’s ineffective for creative tasks but useful for programming, as its limitations help students engage more deeply.
  • Opportunities for Learning: ChatGPT can handle low-level tasks, freeing time for problem-solving, but concerns exist about students missing key skills like debugging and logic generation.
  • Adapting the Classroom: Educators are adjusting assignments to focus on logic and requiring citations for AI-generated work, while also worrying that AI could replace platforms like Stack Overflow.
  • Looking to the Future: Concerns about ChatGPT include reduced critical thinking and increased achievement gaps, as students may rely too much on AI instead of developing independent skills.

Personal Takeaways:

  • Collaborative Research: As the only sociologist on a computer science team, I learned how to effectively contribute to a multidisciplinary research project and bridge different academic perspectives.
  • Applying Sociological Skills: I adapted my sociological methods to a computer science context, using qualitative research to analyze educators’ views on AI in education.
  • Overcoming Challenges: Navigating the technical language and research approaches of a CS environment expanded my ability to communicate sociological insights in a way that resonated with a diverse team.
  • AI in Education Insights: Gained a deeper understanding of AI’s role in education, especially its potential to enhance learning while also presenting challenges around maintaining critical thinking skills and ethical usage of technology.