When to Acknowledge AI Use in Academic Writing
Since the launch of ChatGPT in 2022, artificial intelligence (AI) tools have quickly become part of academia. Though some people were quick to adopt AI use in late 2022, others were cautious and remain hesitant. Nearly four years later, universities are offering trainings on how to use AI ethically as researchers and scholars.
We now have AI tools, such as Co-Pilot, Claude and Gemini, integrated into programs we use to write, email and create presentations/spreadsheets. In addition, a plethora of other AI tools are used for researching, coding, synthesizing sources and creating visuals.
As an editor, I’m seeing more and more generative AI disclosure statements included in scholarly writing. These statements specify ways in which AI has been used/not used as well as what has come from the author/researcher in terms of ideas, thoughts and writing. Among graduate students, researchers and faculty, scholars are still asking:
How do I acknowledge using AI in my academic writing?
The short answer is to be transparent and follow stated policies.
Using AI responsibly in academic writing is about how you use tools, whether that use is permitted and how clearly you communicate their role in producing your work.
Know Your University’s AI Policies
Before using generative AI for academic work, determine what your university, program, department or other governing body permits.
Policies can and do vary considerably. I’ve worked with certain programs where faculty were teaching AI use in the classroom while other departments essentially banned it, unless a faculty member allowed it. So, within the same department, use can be on two ends of the spectrum.
As a scholar, you must do your due diligence to determine whether your document confirms to acceptable standards. You cannot assume that because an AI tool is publicly available or was used in one instance that it can be used at all or for every academic purpose.
Know Publishers’ Policies
University requirements are not the only policies to consider for your academic writing. Academic journals, presses, professional organizations and conference organizers have implemented guidelines and best practices as well. If you’re preparing an article, book chapter, conference paper or other work for publication, review the policies of the organization where you plan to submit your work.
Most, if not all, publishers have established requirements concerning acceptable AI use and disclosure. Many policies are classified as transparent use of AI, just as you need to cite research sources. Know that what’s acceptable in your university setting may not necessarily meet the requirements of a particular publication.
Reviewing relevant or required AI policies should become part of your submission process, just as you check formatting requirements, word limits, citation style and submission instructions.
Questions to Ask Yourself When Using AI
Before using AI, I recommend that you answer the following questions:
Does my university, journal, grant funder or academic press have an official generative AI policy?
Does my program or department have additional guidelines that govern my use as a student or faculty member?
Have rules been established for the document I’m writing?
Do organizations have special requirements for theses, dissertations, publications, conferences papers or writing based on funded research?
Am I required to disclose and/or cite my use of AI?
Where are AI usage policies headed in higher education and how can I protect my scholarship from scrutiny in 5 to 10 years?
When was the last time I reviewed a policy to ensure its updated?
When you can’t locate policies or they’re unclear, find a person to ask before using an AI tool rather than assuming AI use is acceptable.
Disclose How You Use AI in Academic Writing
AI usage disclosure a best practice because transparency builds trust. You should clearly explain which AI tool(s) you used and what you used them to do. This disclosure can appear in the acknowledgments, methodology section, appendix, author’s note, reference list or another designated section. Again, refer to the guidelines of the organization for specifics.
A useful disclosure leaves no ambiguity to the following questions:
What tools or programs did you use?
How did you use it (describe all activities related to researching, planning and writing in detail)?
What role did you retain (identify what you’re responsible for as the author/researcher as you should check and verify what AI did)?
You can even request or seek sample disclosure statements that acknowledge how AI assisted you.
The key part of ethically disclosing AI usage is not simply announcing, “I used AI.” The goal is to provide enough information for a reader to understand AI’s role in your process.
The Importance of Protecting and Cultivating Your Unique Voice
Academic writing is a reflection of your thinking. To get ideas clearly out of your head requires critical thinking to organize information, develop arguments, evaluate sources and communicate ideas well. AI should not do this work for you. While AI disclosure is a responsible academic practice, disclosure does not make every use of AI appropriate.
As a scholar, AI usage should support, not replace, your intellectual work.
A year ago, I returned to the classroom to teach at the undergraduate level. I was saddened by the lack of confidence that students who had become accustomed to relying on AI possessed. These students weren’t clear about their voice or writing style. They thought I wanted writing perfection, but I wanted to know their unique voice and see it reflected in their writing.
We all have a writing style, just as we all have a speaking style. AI makes it difficult to discern that writing style. Why? Because AI writing removes nuanced word usage, sentence structure and writing organization that distinguishes one person from another.
As I received perfect emails, I instantly knew the students hadn’t written the emails because AI has a writing style as well. It is formulaic, repetitive and nearly grammatically perfect. The emails didn’t correspond to the voices I heard in class or the diction used by young adults. The language, punctuation and sentence construction in the emails I read certainly deviated from in-class writing assignments.
So, I caution people against relying on AI too much. Sure, it can write the email, the report and the paper. But ask yourself:
What is having AI write for me costing me?
Confidence?
Growth?
Clarity?
Uniqueness of expression?
Can a reader still identify what I contributed?
Does my writing read like the writing of everyone else who uses AI?
A Case Study: How a Writing Coach Uses AI
I’m an advocate for using resources that make us more efficient, but not at the expense of our growth, especially in the realm of higher education, where learning is supposed to occur.
I use AI, but I’m also extremely clear about my voice. In fact, I turn off AI in my emails and Word software programs because I get tired of it suggesting words or attempting to rearrange my sentences. I construct my sentences in a certain way and intentionally use specific words. I don’t want or need AI to conform my voice to its code or what it predicts should come next.
When I want assistance, I turn AI on or use a paid subscription I have. For instance, AI helps me curate trips and plan meals. Academically, it helps me to create documents based on information I’ve written.
In fact, I used AI in this blog, specifically I utilized my paid subscriptions to ChatGPT to repurpose my writing. For this blog, I input a series of items I wrote and asked it to recommend ways I could use the information as a blog. I gave it instructions on my goals, and it recommended several blogs when I considered using the content for one blog. I often use AI as a strategist, giving it an expert persona to advise me. I agreed with the recommendation. It framed the blog into prior pillars I’ve given it for the work I do. Next, it provided subsections and questions to add to what I input.
In typical generative AI fashion, it provided too much information, repeated itself and offered to assist even more. So, I went to work. I deleted some information, reorganized other items, added sections and edited it all into my voice, my thoughts and my experiences.
I don’t mind AI assisting me in my social media writing or marketing writing. In fact, I’ve learned how to better chunk my ideas and bullet main points to make it more digestible by noting how it presents information. And, I incorporate that approach into my email writing and marketing materials (I also can tell people who use AI in their materials because of my familiarity with AI’s style). Furthermore, I’m not strong at creating spreadsheets and visuals, so I use Co-Pilot to take information I’ve written to create figures, graphs and spreadsheets, but I closely review and edit everything it produces.
I’ve become comfortable with these uses of AI. Still, I am deliberate about preserving my voice. Plus, I have a hard line that AI will not cross in my writing and planning. I do not use it to assist with the PowerPoints I train from or the books I write.
Currently, I’m working on a book that has been in the writing process for more than two years. People often ask me:
“Why don’t you use AI to write your book?”
My answer: I don’t want to short circuit my thinking. I want to grapple with my ideas, think about what I want to say, clarify my thoughts, revise my writing repeatedly and think about if what I’ve written is effective. I challenge myself to reach my audience with the right examples, structure and information.
In this process, the final version is nothing like the first version. My ability to communicate grows. My ability to critically think improves. My writing becomes more refined. Yes, it takes me years to write a book, but what I gain in the process is invaluable.
My authentic voice is cultivated, and my thinking is sharpened.
This type of growth would not occur if I simply prompted AI, based on my purpose, audience and an outline, to generate the contents of a book for me. I’m sure it would generate good content, but the content wouldn’t be mine. The style wouldn’t be mine. Most importantly, my thinking would be absent from the book, which means my voice would be absent too.
Simply put, AI should not overshadow or augment your voice. As a scholar, thinker and researcher, you are responsible for your work, not just the accuracy of the information, but the thinking behind it. AI tools can support your thinking and writing, but you have an intellectual responsibility to generate the content.
Generative AI is constantly changing, and academic policies will continue to evolve. As an academic writer, you don’t need to predict every development, but you need to develop scholarly practices that support an ethical, transparent use of AI that facilitates the development and presentation of your unique voice.
AI can absolutely support parts of the academic writing process. However, it should not make it unclear who did the scholarly work.
Comments