Introduction to leveraging GenAI for learning Academic English

Last updated:

Do you know…​
- 84.3% of students use ChatGPT at least once a week as a study/research tool,
​- 86.5% of students believe that AI positively affected their learning,
- 91.7% intend to continue using AI as a learning tool in the future.
(Data retrieved from He et al. (2025))

Introduction

​Generative Artificial Intelligence (AI) is rapidly transforming the way we work, study and even perform fundamental tasks like thinking and reasoning. AI technology is on the path of becoming an integral part of teaching and learning, and students’ ability to use AI effectively will likely affect their academic performance and future development. However, many students are unaware of the university AI policy, causing confusion, reluctance and even potential misuse (Barrett & Pack, 2023; Chan & Hu, 2023). 
  • How do generative AI tools work?
  • How can students make use of AI for self-learning?
  • How do university policies regulate the use of AI?
​In this guide on using AI for learning, you will learn more about the university’s AI policies and regulations, basic functionalities and limitations of generative AI, your ethical and academic responsibility as students as well as ways to incorporate AI into your studies. 

​Generative AI – A Quick Breakdown

Q: What is Generative AI? ​
  • A: Artificial Intelligence (AI) are…
  • Technologies that can process and handle intricate and cognitively demanding tasks by simulating human intelligence using computer programs and algorithms (Mao et al., 2024; Popenici, 2023). 
​Generative AI, a sub-category of AI, are capable of generating highly coherent, complex and diverse texts or artefacts that are virtually indistinguishable from those created by a human (Chan, 2023). 
Q: How does Generative AI work?
  • A: Generative AI models predominantly use machine learning and deep learning techniques to mimic patterns from existing datasets and to generate original data samples, which can come in various forms, ranging from texts, images, and sounds to video and even computer codes (Chan & Hu, 2023). 
​Large Language Models (LLMs), which are trained on a large corpus of natural language text, are currently the most widely used type of generative AI (Chan & Hu, 2023). Some commonly known examples of generative AI models include ChatGPT (OpenAI), Gemini (Google), Claude (Anthropic), Grok (X), DeepSeek etc. 

Some key technological features of GenAI: ​

Tokenization​
Tokens are the basic unit of data that LLMs process (e.g. words/sub-words), the number of tokens that a model handles at a time affect both cost and performance.
Attention mechanism
LLMs use the attention mechanism to weigh the importance of tokens relative to each other, allowing them to prioritize on key information and generate coherent and context-relevant output.
Context window​
LLMs have a fixed amount of working memory that affects the number of tokens that they can process. Models with larger context window can recall prior conversation or data better, making them better at handling more complex tasks.
Q: Is generative AI truly intelligent? 
​A: While generative AI is certainly powerful in many ways, such as generating highly sophisticated texts, programming advanced codes and understanding complex commands, the current iterations of generative AI models rely on mechanisms that are fundamentally distinct from human intelligence. 
​Some scholars describe the nature of generative AI as “deceptive”, as they do not engage in thinking: 
​“It is deceiving to say, dangerous to believe, that artificial intelligence is… intelligent. There is no creativity, no critical thinking, no depth or wisdom in what generative AI gives users after a prompt: it is just plausible text with good syntax and grammar, and this is all that it is.” (Popenici, 2023, p. 383)
​It is also important to stress that AI responses are “probabilistic” in nature, meaning that they are always prone to making errors and providing illogical answers: 
​“Generative AI such as ChatGPT does not work in the way that a calculator works. A calculator actually performs the calculations required in order to reach an answer; this is important. Generative AI does not perform calculations, it does not go through the learning, it does not engage in thinking. ChatGPT and similar tools make predictions; they guess.” (Lodge et al., 2023, p. 3)

Using AI in your course assignments

​How much you can use GenAI in your assignments can vary significantly depending on your course/department. Some fields may encourage the integration of AI tools to enhance research and improve efficiency, while others might have stricter guidelines regarding originality and authorship.

​For example, in technical disciplines, using GenAI for data analysis or coding assistance might be welcomed, whereas in humanities or social sciences, reliance on such tools could be viewed as compromising the integrity of your work. Additionally, specific assignments may have their own guidelines regarding GenAI usage, so it is important to understand the expectations for each task. Here are some examples that show the varying extent to which AI use is permitted in different assignments (Flinders University, n.d.):
Click to See Examples
Extent of AI Usage
Examples
AI in the planning stages of a task
The use of AI for an initial scan of the literature
The use of AI to identify the sections and elements needed for a text type
AI as a core part of the task
The use of AI to improve transfer of knowledge skills
The use of AI to learn evaluation, critical thinking and reflection skills
The use of AI for divergent thinking and the generation of ideas 
AI for self-testing
The use of AI to generate quiz questions for themselves on the content to facilitate revision
The use of AI to explain why an answer/ solution is correct or what other options may be
AI as a copyediting tool
The use of AI to make edits to their own work
The use of AI to identify and improve story style and tone

Potential of Using AI in EAP Learning

​In the context of EAP, generative AI has great potential for supporting your learning. For example, AI can help with…
Click to see examples
Task 
Example Prompt
Content Development
Suggesting potential topics outlines or resources.
E.g. “Write an essay outline for the topic – AI and research ethics.” ​
Corpus Search
Searching a word or a sentence within a corpus.
E.g. “Find a sentence in this article that uses the word *critical* or similar.”
Text Modification
Adjusting writing to fit a specific genre, tone, voice, or formality.
E.g. “Rewrite this sentence in formal tone: I swear it is pretty obvious that hot weather causes bad mood”.
Feedback and Revision
Analysing learner’s writing errors and mistakes.
E.g. “Analyse the strengths and weaknesses of this essay: *Insert writing*.”
Response to Questions 
Providing learners with answers to specific questions.
E.g. “What is the meaning of the word “retrospect”?” ​
Sentence Generation 
Generation of templates or sentence starters.
E.g. “Write a sentence introducing the benefits of AI.”
Vocabulary 
Providing explanations or synonyms for unfamiliar words.
E.g. “Suggest a list of synonyms for the word *significance*.”
Grammar Check/ Editing
Checking the grammar of your writing.
E.g. “Check the grammar of this paragraph *insert writing*.”
Paraphrasing
Paraphrase a piece of writing.
E.g. “Paraphrase this sentence *insert writing*.
Summary
Summarizing key ideas of texts to meet your needs.
E.g. “Summarize the key findings of this article in 200 words.”
​Table adapted from Feng Teng (2024, p. 53)
​Students are encouraged to be creative and experiment with innovative ways to help them improve writing, speaking and other aspects of EAP. However, the use of generative AI also poses potential risks to students, especially in the context of EAP learning.

Risks of Using AI

The most prominent concern for students using AI is their potential overreliance on the tools which could undermine students’ development of writing competence (Chan & Hu, 2023). Concerning language proficiency, AI analysis of student writing may not accurately reflect their actual language skills. As a result, students might misjudge the level of their abilities or knowledge, over or understating their actual strengths and weaknesses, which could lead to reduced learning motivation (Wang et al., 2023).
What are the risks?
  • Misinformation and inaccuracies (Hallucination)
  • Biases and misconceptions
  • Academic misconduct (e.g. plagiarism)
  • Privacy and Data Security
  • Undermining human values
  • And more… 
​It is crucial for students to take the potential risks that accompany generative AI tools into account and exercise appropriate care when using them as learning and assessment aids. 
WHO WE ARE
VIDEOS
EVENTS
MATERIALS
Copyright © 2026. All Rights Reserved.