Welcome

Welcome to this online guidebook designed to help English Language Learners (ELLs), understand how to ethically use AI in your academic writing. This guidebook provides an overview of what AI is, how it works, its benefits and limitations, and—most importantly—how you can use AI ethically in your academic writing while maintaining academic integrity. Although this guide is primarily designed for ELLs studying at U.S. universities and writing for academic purposes, others who want to learn how to use AI ethically in academic writing may also find it useful. Keep in mind that this guide is not comprehensive; it's a general overview, and you should consult additional resources to fully understand the ins and outs of using AI in academic writing.

Understanding Generative AI

When you hear people talk about using AI for schoolwork, they are often referring to generative AI models like ChatGPT or Gemini. The key word here is "model." This means that, like all models, AI is a representation of reality and therefore can produce erroneous outputs or fail to understand the context in which it is used. This is a key fact to keep in mind when you use any AI. Think about this as you read the information below.

How Generative AI Works

In simple terms, generative AI analyzes massive amounts of data to generate a potential answer to a prompt you provide. AI-produced answers are based on patterns learned from vast datasets. Generative AI connects different pieces of data, often resulting in multiple answers to the same prompt. For example, according to Daumé (2023), if you ask an AI model to complete the sentence "I went to the Zoo to see the ___", it will analyze the data it has access to to make connections and create possible answers like animals, giraffes, or penguins. If you want to learn more about how AI works, this site has additional information

Using AI in Academic Writing

AI tools can be incredibly helpful in your writing process, assisting with tasks like brainstorming ideas, improving grammar and vocabulary, offering research assistance, learning about citations, summarizing and paraphrasing, and editing. Grammar and vocabulary help, as well as paraphrasing and summarizing, can be especially beneficial for ELLs. However, you should keep in mind that AI has limitations (see the section below), and the English used to train AI may only represent certain types of English. For instance, much of the information online is written in U.S. standard English. AI often struggles to represent the complexity of human language. As mentioned above, AI makes predictions based on the data it has been given, meaning that the output often repeats prescribed norms.

Maintaining Your Own Voice

Given this, you must be careful not to over-rely on AI models, as this could hinder the development of your personal writing style or "voice." U.S. academic culture strongly values originality, and your professors will expect you to create new and unique ideas. Over-reliance on AI may prevent you from achieving this, as AI often replicates existing ideas. Additionally, as mentioned above, the language used by AI often reflects one kind of English and overlooks the complexity of human language.

Using AI ethically means ensuring that the final content is authentically yours. You should use AI for tasks like brainstorming, guidance, and refinement rather than creating the entire content. Your ideas should form the core of your work. Think of AI as an assistant to help you in the process, not to do the work for you.

Limitations of AI

As we've seen, AI tools are not perfect and can sometimes produce biased content. Consider current inequities in society. For example, racial bias in law enforcement could be replicated by AI tools if they are trained on biased data, leading to biased outputs. Moreover, AI makes connections between content that has already been made; thus, over-reliance on AI could lead to a lack of new thought and development, particularly in your writing skills. Because AI works with what has already been produced, it can be challenging to develop your own style or "voice" when using AI alone.

Additionally, Shilton (2023) points out two other concerns: accuracy, and privacy & security. First, AI is not necessarily accurate in the way you might expect. It provides an answer based on the data it has been fed, which means that the answer may not be correct if the data is flawed. Second, generative AI tools may learn from users, meaning they could use the information you provide to further train themselves. Be cautious about what you share with AI. You don't want sensitive or personal information to become part of its dataset and potentially shared with other users.

Academic Integrity and Plagiarism

Understanding Plagiarism

In U.S. academic culture, many students struggle with plagiarism, which is using someone else's work or ideas without proper attribution. This includes copying text, ideas, or even AI-generated content without acknowledgment (acknowledgment means proper citation). While the purpose of this guide is not to teach you how to cite, you should know that you must use some type of citation to show your professor that you've used AI for your work. Common citation styles include APA, MLA, and Chicago. To learn more about this topic, you might take a look here.

If you don't cite the AI model you use, you risk failing an assignment or losing your student status. You should become familiar with the academic integrity standards at your university. Be sure to carefully review them, as standards vary by both university and culture. Familiarize yourself with the unique rules at your institution. For instance, the University of Maryland has a detailed code outlined here. Finally, note different professors may have different standards for ethical AI use. You should thus always ask each professor how and when AI can be used on assignmetns in every course.

Fact-Checking AI-Generated Content

Why Fact-Check?

As you've learned, AI tools often produce incorrect or misleading information. No matter what you do with AI, you should always double-check that the information is correct. One simple way to do this is through a strategy called lateral reading. Lateral reading involves opening multiple new tabs in your web browser and searching for the information you’ve been given to ensure its accuracy (Digital Inquiry Group, 2020). This website has great information on how to read laterally, including videos demonstrating the strategy. You want to see if other sources provide the same information. Keep in mind that you should be looking at verified sources. If you’re not sure how to do this, you might want to take a look at this site.

Methods for Fact-Checking

• Cross-reference with multiple reliable sources.
• Use academic databases like Google Scholar and JSTOR.
• Critically evaluate the credibility of sources.
• Utilize fact-checking websites like Snopes and FactCheck.org.

Example Exercises Using AI

At this point, you might be wondering what kinds of activities constitute ethical AI use. As mentioned above, think of AI as an assistant. While it can help you in the process, it shouldn't do the work for you. Below, you'll find examples of ways you can use AI ethically in your writing process. These examples have been adapted from a conversation with ChatGPT (OpeanAI, 2024). Keep in mind that these are only examples, and that AI policies may vary by instructor. Always consult your instructor before using AI to ask them what they consider unethical use!