Skip to main content Skip to navigation

Potential Societal Biases of ChatGPT in Higher Education: A Scoping Review

Project Overview

The document explores the impact of generative artificial intelligence (GAI), with a focus on ChatGPT, in the realm of higher education, revealing both its transformative potential and ethical challenges. It identifies key applications of GAI, such as personalized learning, automated tutoring, and administrative efficiency, which can enhance teaching and learning experiences. However, concerns arise regarding the biases inherent in GAI models, stemming from the datasets used for training, which can adversely affect fairness and inclusivity in educational settings. The findings underscore the urgency for comprehensive research into these biases, particularly in diverse fields beyond the traditional focus on medicine and engineering. Furthermore, the document advocates for the establishment of ethical guidelines to govern the deployment of GAI technologies in education, ensuring they are utilized responsibly and equitably. Overall, while GAI holds significant promise for improving educational outcomes, careful consideration of its ethical implications is crucial for its effective integration into educational practices.

Key Applications

ChatGPT for Educational Applications

Context: Various educational activities including teaching, learning, research, language tutoring, and administrative processes in higher education institutions, targeting students, faculty, staff, and researchers.

Implementation: ChatGPT is integrated into educational processes across multiple contexts, assisting in creating teaching materials, analyzing student data, enhancing accessibility for students with disabilities, and streamlining research analysis. It is also utilized in admissions and decision-making processes to improve efficiency and support language learning.

Outcomes: Generates teaching materials, enhances language acquisition, identifies learning patterns, enhances accessibility for students with disabilities, streamlines research processes, and improves efficiency in processing admissions applications. However, there are inherent risks of misinformation, biases in outputs, and potential impacts on academic integrity.

Challenges: Ethical concerns regarding biases in AI outputs, potential decline in academic integrity, risk of misinformation through spurious references, biased judgments based on personal data affecting admissions decisions, and the reinforcement of biases in educational content design.

Implementation Barriers

Ethical

Concerns regarding biases in AI outputs that may reinforce existing societal inequalities, including inherent biases in training data that can lead to biased outputs in GAI applications.

Proposed Solutions: Develop clear ethical guidelines for the use of GAI in educational contexts, including protocols for identifying and addressing biases. Incorporate diverse and representative datasets into GAI training processes to mitigate biases.

Cultural

Lack of discussion about biases in non-English and multilingual educational settings.

Proposed Solutions: Encourage research on GAI applications in diverse cultural contexts and develop culturally sensitive AI systems.

Project Team

Ming Li

Researcher

Ariunaa Enkhtur

Researcher

Beverley Anne Yamamoto

Researcher

Fei Cheng

Researcher

Lilan Chen

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Ming Li, Ariunaa Enkhtur, Beverley Anne Yamamoto, Fei Cheng, Lilan Chen

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies