Abstract
This chapter presents a comprehensive literature review to identify the challenges and limitations of using generative artificial intelligence (GAI) in education. As a result of screening seven major citation databases, 476 studies were reached. Analysis was carried out on 25 studies selected according to the inclusion and exclusion criteria. Results showed that research on using GAI in education is mostly conducted at the higher education level. The number of studies focusing on lower levels of education is quite low. The challenges and limitations of artificial intelligence are more about general education rather than focusing on a specific discipline. ChatGPT was the most investigated GAI tool. The challenges and limitations of using GAI in education are grouped under five factors: ethics and safety; educational implementations; assessment and evaluation; equity and access; quality control and expertise.
TopIntroduction
Generative Artificial Intelligence (GAI) has gained significant attention in education, presenting both opportunities and challenges. The integration of GAI in education has the potential to revolutionize teaching and learning practices, but it also raises concerns regarding ethical use, equality, and the role of educators. Several studies have explored the limitations and implications of GAI in education. From an ethical perspective, Lin (2023) argued that there is no need to restrict the scope or nature of GAI assistance as long as its use is transparently disclosed. However, detecting fake research poses a critical challenge, which can be addressed through open science practices such as transparent peer review and data sharing (Lin, 2023). Implementing GAI in education requires careful consideration of ethical considerations, attribution of credit for GAI-driven discoveries, and advancing educational goals (Alasadi & Baiz, 2023).
The proliferation of GAI tools like GPT-4 and Open Assistant presents a paradigm shift in information acquisition and learning, posing substantial challenges for traditional teaching approaches and the role of educators (Walczak & Cellary, 2023). GAI tools, such as ChatGPT, have the potential to impact computer science education. Challenges for computer science education brought on by GAI tools include potential competition for computer science educators and the need to respond to possible impacts (Van Slyke et al., 2023). However, the consequences of GAI in education can vary, from having little impact to serving as competition for educators (Van Slyke et al., 2023). While GAI raises concerns about plagiarism detection, it also presents opportunities for educators to leverage GAI to build supportive learning environments (Eager & Brunton, 2023). GAI-augmented teaching and learning practices in higher education offer productive affordances but also come with challenges that need to be addressed (Eager & Brunton, 2023). Integrating generative chatbots in higher education can enhance the educational experience and facilitate effective implementation (Ilieva et al., 2023). However, it is crucial to explore best practices and strategies for utilizing GAI for educational purposes (Michel-Villarreal et al., 2023).
There are numerous studies on the potential benefits of using GAI in education (Bahroun et al., 2023; Chiu, 2023; Lo, 2023; Mao et al., 2023; Montenegro-Rueda et al., 2023; Popovici, 2023; Pradana et al., 2023; Preiksaitis & Rose, 2023; Sallam, 2023; Zhang & Tur, 2023). Bahroun et al. (2023) systematically synthesized 207 articles to provide a comprehensive analysis of the use of GAI in education. The research examined the transformative impact of GAI in specific educational fields, including medical and engineering education. Chiu (2023) investigated how GAI changed school education using qualitative methods. He revealed the views of 88 school teachers and administrators on using GAI in education within the categories of learning, teaching, assessment, and management. Lo (2023) conducted content analysis on 50 articles to enrich our understanding of using GAI in education. Mao et al. (2023) examined the potential opportunities of using GAI to summarize the use of AI in education. Montenegro-Rueda et al. (2023) presented an analysis of the impact of applying the GAI tool ChatGPT in education by conducting a systematic literature review in three leading scientific databases in the world of education (Web of Science, Scopus, and Google Scholar). Popovici (2023) used ChatGPT to solve all coding assignments given during the semester in a Functional Programming course to explore the capabilities of GAI and evaluate its value for educational purposes. Pradana et al. (2023) reviewed existing research on using the GAI tool ChatGPT in education using bibliometric analysis and systematic literature review. Preiksaitis and Rose (2023) synthesized the potential opportunities of GAI in medical education. Sallam (2023), using the systematic review method, investigated the potential utility of the GAI tool ChatGPT in health services education, research, and practice. Zhang and Tur (2023) used the systematic review method to investigate using ChatGPT, a GAI tool, in educational settings from kindergarten to grade 12 (K-12). Unlike the studies in the literature, this study focused on the challenges and limitations of using GAI in education.
Key Terms in this Chapter
ChatGPT: It is an artificial intelligence model used in natural language processing. It is a language model developed by OpenAI that can generate realistic and meaningful answers in text-based conversations.
Database: It is a database of citations of scholarly articles, books, and other academic sources, helping researchers track, cite, and search literature.
Measurement and Evaluation: It is a concept in education that enables the systematic and objective determination, analysis, and interpretation of a process or a student’s performance, abilities, or achievements.
GAI Tools: Generative AI Tools refers to AI-based tools that allow users to generate text, images, music, and similar content automatically, often designed to assist in the creative process.
Open Assistant: It is an AI assistant that can be freely developed and personalized.
GAI: Generative AI is a concept that refers to artificial intelligence models that can generate new and original content by learning from data sets.
BERT: BERT (Bidirectional Encoder Representations from Transformers) is a deep learning model used in natural language processing.
Ethics: Ethics is a philosophical concept that deals with moral principles and rules of behavior as a field in which right and wrong, good and evil, just and unjust, and honest and fraudulent behaviors are examined and evaluated.