Abstract
The chapter explores the transformative potential and challenges of integrating large language models (LLMs) into higher education. It highlights the opportunities AI presents for enhancing academic literacy, writing, and pedagogy, while also acknowledging the risks to traditional educational values and practices. It proposes a framework, developed with the guidance of academic literacy, information literacy, digital literacy, and academic integrity, aimed at leveraging AI's capabilities to support academic success without undermining foundational skills. The discussion extends to the implications of AI in the South African educational context, addressing the digital divide and advocating for equitable access to technology. This chapter encapsulates the essence of the proposed proactive framework for navigating AI's impact on academia, focusing on adaptation, critical engagement, and the cultivation of an advanced form of academic literacy that integrates AI technologies responsibly.
TopIntroduction
With the emergence of artificial intelligence (AI) technologies, particularly Large Language Models (LLMs), in late 2022, academic literacy as a practice and pedagogy is on the cusp of transforming (Grassini, 2023). In the face of these advancements, higher education institutions (HEIs) are confronted with what Gimpel et al. (2023: 28) describe as the “educator’s dilemma” where decisions need to be made on either banning such tools or fostering their responsible usage. The core of contention is the rapid content generation capability of LLMs (Elkhatat et al., 2023; Chen et al., 2024), which presents a contrast to traditional research and education practices of HEIs. This situation necessitates careful consideration and strategic planning for the integration and adaptation of AI technologies, especially within a South African context.
This chapter, therefore, proposes an integrative framework designed to harness AI's potential while anchoring educational values in the relevant digital, academic, and information literacies. Simultaneously, it emphasizes alignment with the bedrock principle of academic integrity. Our aim is not merely to equip learners with the tools of the future but to ensure these tools amplify, rather than eclipse, the functional and foundational academic skills requisite for success in higher education. It is, however, important to underscore the progressive and anticipatory nature of our endeavor. This chapter does not claim to provide conclusive research outcomes; instead, it aims to offer a proactive framework for navigating the burgeoning AI revolution in academia, particularly within the distinct South African literacy context. In doing so, our aim is not just to keep pace with rapid developments but also to develop a constructive critique that influences the trajectory of educational development, ensuring that our institutions remain adaptable, critical, relevant, and forward-thinking. Our approach, therefore, is more than conceptual; it's a commitment to cultivating an advanced form of existing literacies that leverages the positive aspects of LLMs. We provide educators with actionable insights for integrating these technologies into assessment design and curriculum development, thereby enriching the academic experience, and preparing students for a future intertwined with AI advancements.
Key Terms in this Chapter
Large Language Models (LLMs): Applications trained on extensive language data that predict and generate text based on input prompts in conversational manner.
Hallucination (in AI Context): A phenomenon associated with LLMs whereby these models unintentionally fabricate information, posing challenges to the integrity of information.
Digital Divide: The disparity between those with convenient access to information and communication technology (ICT) tools and those lacking such access or the skills to utilise them.
Prompt Engineering: The skill of formulating queries to elicit relevant and appropriate responses from LLMs models.
Information Literacy: Skills required by individuals to recognise when information is needed, where to locate, evaluate, and use effectively in line with recognised standards.
Academic Integrity Literacy: Knowledge of academic integrity values, behaviours, and ethical decision; including, the skills and actions necessary for the ethical use and representation of information in academia.
Generative Pre-Trained Transformer (GPT): “Generative” is the capability of LLMs to generate text, which is contextually relevant to the inputs, “pre-trained” signifies that the model has undergone initial training on vast language datasets enabling it to produce human-like text, and “Transformer” is a type of model that weigh the significance of different words in the input data.
Artificial Intelligence Literacy: The knowledge and understanding of AI necessary for individuals to participate in the broader discourse around AI and make informed decisions about its use and implications.
Linguistic Capital: Rooted in Bourdieu's theory, where linguistic proficiency plays a critical role in academic success, especially in a socio-linguistically diverse context like South Africa.
Academic Literacy: The ability of students to comprehend, navigate, and apply conventions of academic discourse.
Digital Literacy: The ability to understand and use information from various sources presented through digital devices, with specific emphasis on the skills to extract, engage, integrate, and employ information from digital platforms effectively.
Academic Integrity: A commitment to five fundamental values of academic conduct: honesty, trust, fairness, respect, responsibility, and courage (International Center for Academic Integrity [ICAI], 2021).