What's my line?
Exploring how we figure out what is our line when it comes to using AI in scholarship and research.
The following is formed from a recent workshop that I did for Paperpal around how to think about what is appropriate and inappropriate use of AI in writing and research. It was named “Writing with AI: Making Informed, Defensible Choices Before You Submit”. The session was a workshop and had several engaged elements so what I’m sharing here is an adapted essay resulting from the transcript of what I said and some reflections on some of the outputs and such from the participants.
And of course, here are the Slides and Resources. In fact, you can even watch the video too.
The session included a mix of independent scholars, faculty, and students. They came with the hopes of figuring out what the right choices were to make in their writing and research. The problem is, of course, that there are no perfect answers. We’re all still figuring this out. What I hoped for that session (and this post) is to guide the conversation to first recognize that many of the questions have a less-than-exciting answer, “it depends” and a significant part of what it depends on is who you are as a writer and scholar.
It sounds very lawyerly. It can sound like I’m hedging my bets, but this is the reality of where we are. Yet within that murkiness, you can actually walk away with some clarity about where you stand, about what is valuable to you, and that you have a sense of knowing where to go for your questions and finding answers when you figure out what the exact context is that you’re working in.
First, take an inventory of where you are now.
These polls offer a moment to see there are similarities and differences. That we’re all in different places while grappling with the same set of questions. That confusion is the standard bearer of this moment.
The polls provide you, the reader, an opportunity to more consciously think about where you are, if you haven’t, or even to consider whether your answers have changed. Would you have answered the same way 6 months ago? 2 years ago?
One of the observations by participants in the workshop, and it’s one that I think about a lot in this space, is how to prevent AI from taking over my writing style. How do I make sure that I’m not read as more AI slop (to the degree that I can’t do that, knowing folks are going to think what they think). For those of us who see ourselves as writers, there are a lot of questions about our voice.
Given how much I work with AI, I continually have this thought: “Gosh, everybody knows I work with AI. Does this [piece of writing] sound too much like AI, even when I haven’t used any AI whatsoever?” What does that mean for how they view the work? What does it mean when I suspect or can see (what I believe to be) the hallmarks of AI writing in other people’s writing? It’s part of the reason I’m also thinking about documenting and disclosure in the education space beyond the classroom.
In asking these questions, I also want us to avoid too much judgment. We live in an age where judgments are easier to cast than ever, and judgments dominate our attention economy. This is a tremendously complex technology that we’re all still figuring out our own use cases for. In that line, it’s going to be different for different people with different purposes. I keep hoping we get better at developing our own position statements or treating our AI use as a methodology framework.
What’s my line?
Still, the question I keep coming back to when folks are looking for clearer guidance is the title of this piece. We should all be answering this. Maybe some of us can come up with a global answer about where they do and don’t use AI. I suspect more of us will need to just think about it contextually. In any use case, we should ask where is my line in using AI and be able to also answer why. But that needs to be an honest why. A reflective why that names things we might not be comfortable saying about ourselves.
This is a rapidly changing technology. In its ubiquitous form, it’s been around for three and a half years, and what it can do has changed over that time. Hallucinations are real, and we also know in certain ways they’re significantly reduced by certain practices. It also depends on the tasks we’re asking, the models we use, and the tools we choose. How you get results from NotebookLM versus ChatGPT versus Claude is going to create different experiences about what these things can do, how accurate they are, etc.
AI has produced some new ways of doing writing and research. I can do things that I might not have been able to do before, or find new workflows. One of my favorite workflows for writing is to identify what it is I’m going to write, what the purpose is, what the focus is, and write a good prompt for an AI tool of my choice and set it up to interview me. Through that, I can figure out what I know, what I need to know, and what I need to do next. The AI questions and challenges my thinking in ways that I’m not thinking about, which leads me to pursue other avenues of thought. That new workflow has been much more helpful than staring at that blank white screen, and I say that as a professional writer of 20+ years.
But the challenge that many of us are dealing with is that our fields and disciplines are deliberately moving. They do not move fast, and that is sound for what they do. However, nearly 4 years into AI, we’re all feeling the tension of these ubiquitous tools, and what our fields expect or even understand about where they fit.
Of course, within all of this, there is appropriate skepticism of the tools and what they can and can’t do. And we want to hold on to that skepticism. Like the good researchers and scientists we aspire to be, we have to actually be more deliberate. We have to challenge all of the things that we see coming out, and not just the things that reinforce our views.
We will have to engage, experiment, and advocate in our fields to recognize that AI, just like the internet, is likely to change all of our disciplines and different kinds of work in all of our disciplines. Our responsibility is to help figure out what feels appropriate. What do we hold on to? What do we let go of?
Before we can answer that for our disciplines, we have to answer it for ourselves.
So all of this is to say: when I talk about where’s my line, it’s really thinking about you as an individual and what it is that you value, what is your purpose, your drive. Why are you a scholar? What is it that’s driving you to do research? What makes sense for how you want to show up as a writer, researcher, or scholar? We each need to come up with our own individual policy, our own internal sense of “here’s what I do and why” as it relates to AI.
I could spend days providing tips, tactics, do’s, and don’ts. But none of those things matter unless you have your own North Star.
All of our questions about AI are really about our own ethical practice, and recognizing where they do and don’t come into contact, in contrast, with different publishing spaces or different ways you want to show up as a professional doing intellectual work.
At this juncture in the talk, I had them engage in a Padlet activity. For ease of recreation, I had five columns: Early-stage thinking, planning & structure, draft development, revision & editing, and source-related uses. Under each, they were to articulate where is their personal line. Not what they think they are supposed but what they feel is right for them. The goal was that given the anonymity of the Padlet, they could share where they honestly were.
The most striking aspect about the activity was that there were 310 responses across those 5 areas. The majority of those responses seemed to focus less on whether AI touches academic writing and more on whether AI originates (authors?) the writing. Revision, editing, grammar, summarization of already-read material, and adversarial stress-testing of one’s own ideas were approaches that participants leaned into. Concerns about draft generation were evident with multiple participants’ use of explicit boundary language (”my line,” “I don’t dare”). Taken together, these contributions seemed to indicate an unspoken principle among this group: AI may refine, interrogate, or compress content that the author has already engaged with, but should not be the workhorse of thinking itself.
What is my practice?
I want to get more specific and practical in my own practice here. Each are questions we should each be asking ourselves as we explore AI in our work.
Where do I begin figuring out AI usage?
It starts with my personal goal and intention. Am I writing something that’s supposed to represent me? Am I writing something meaningful to me, that I feel I have something to offer or put out to the world? Or am I writing something that has less emphasis on that?
For instance, I have no problems having AI create instructions, and I edit them. That’s reasonable. I don’t want to be known as the person who writes great instructions. Not because that’s not important but I don’t believe it’s what I offer the world in terms of my abilities and skills. I do have a problem using AI to an essay like this. It may help me clean up the transcription that it comes from, but I still had to put hours into shaping it into something that feels aligned with my voice and intentions.
What drives you?
There’s a deeper question to consider that’s deeply personal and different for all of us. What is your purpose and value in showing up in academia to do research and writing? I encourage you to take the time to actually think this through or write this out. Many of us find our way into our disciplines out of passion, out of love, out of curiosity. We want to hold that in front of ourselves. Many of those feelings come from our own internal experiences. When we make connections, when we have those mental explosions of ideas, when we put our ideas together in a coherent manner. That’s a challenging question as we approach AI, and it requires a bit of introspection before looking at AI. It may sound cliché, but devising your own mission and values statement as it relates to your scholarly identity could be incredibly valuable (and putting it right next to where you work for a reminder).
Would you be comfortable with people seeing your chatlogs?
It’s old advice and one from a mentor of mine when I worked in a state institution. He said, don’t put anything in a work email that you wouldn’t be comfortable seeing blasted on the front page of the New York Times. I think that’s the best practical advice with AI. Run a thought experiment with each usage or chat. “If this chatlog was shared in tandem with my work, would people determine that my work wasn’t sufficient enough to claim as ‘mine’?”
You may not have perfect answers for that. But you need to have that conversation with yourself and then press yourself further and ask, if you are uncomfortable with it, why? Why would it feel like a problem if somebody were to see my chat logs that were used and helped build this project?
An important consideration here is just the idea that, can you authentically stand behind everything in that piece that was created? Can you justify every piece of knowledge, insight, and angle? Because if you’re putting something forward with your name on it, that’s what you have to do. Authorship means “this fully represents me.” And if you can’t account everything in that work, you’re setting yourself up for some possible negative outcomes.
I say this and I also know—we all see AI being used in a lot of contexts whether people are or aren’t supposed to do so (often without articulating or making transparent their use). There’s increasingly signs of improper use showing up in our work. One example is the research emerging around “tortured phrases”—new terms that arrive in research that have no predecessors and become evident they were generated by AI. The question we need to grapple with isn’t what others are doing; it’s what am I doing. What happens when I have to defend ideas, words, research that I can’t whether it’s because I don’t understand what I put out there or that the AI made up?
How do I tend to my usage?
If you are like me, you’re often operating and engaged in several different AI tools in a given day. So it can be really challenging to think about: well, where did I use it, and how did I use it, and all of those things. And so there are a couple of different approaches.
One approach I have in my head is to start thinking about AI usage as a methodology. Increasingly, in things that I’m writing and publishing in different journals and articles, I articulate my usage. If we start to talk about it as a methodology, it can help us think about what to track, what to do or not do, and how to attend to that in our outputs. More recently, I’m trying to think in advance of what it would mean to intentionally lay out how I plan to use it, describe it, and come back after to verify that was my approach.
A simpler version to start with is just remembering to put links to chatlogs at the bottom of a document you are working on. Sometimes, I might have 3-5 different chatlogs for different reasons. Creating a space in a document to track that so you can revisit and see what’s in your work and what’s from AI can be really helpful at the end.
How do I keep up?
Again, the resources tab on the document is a good start. Going to publishers’ websites also can be helpful for specific publications contemplation. But you can also just check out AI policies of journals and publications you follow (often under author guidelines) that can help you get a sense of what are the current expectations and watch them change (or not) over time. Along those lines, also check out the works that you’re reading to see if there are AI usage statements. What do they say?
A deeper approach is to gather your disciplinary insights about where AI is situated in the writing and research process. Even before generative AI came out, some disciplines that were already having this discussion; others are just starting. This can be your disciplinary associations. Many have started this conversation, whether that’s putting out a resource document, a statement, or a set of values and practices. Like anything we’re researching, getting a sense of where the conversation is in a discipline is a smart start.
Many academic communities already exist where these questions are being explored. There are listservs and group communities across Google, Facebook, and LinkedIn.
Those will tell you what disciplines and journals are expecting. Yet another way to just keep up is to have conversations with other people to learn what they are doing, figuring out, determining their line, etc. In those conversations with peers, colleagues, etc, you can engage with them about what you see and also learn what feels right for them, why, and what that means for how you want to show up as scholar.
An extension of that is just building a community of practice. Find people doing similar or adjacent. These may be people you write with, provide feedback to and from, or just have conversations about what it means to be a scholar, researcher, faculty member, or graduate student at this time. These tools are not only present, but they continue to change (not even touching agentic AI in this conversation), and navigating that is best done in conversation with others.
This can be a larger group, and it can be just 2-3 people meeting regularly to share work and questions. It can be a chat thread. It can be a WhatsApp group. It can be a lot of different things. But really making sure that you have other people that you can test your ideas with, that you can build trust and figure out new perspectives that aren’t going to be yours, aren’t going to be the AI’s.
Stuck in the middle
It is okay to be confused. There’s a lot to be confused about. At the end of the day, it’s important for all of us to articulate our confusion and to ask for help. Help can be another opinion, getting clarity, or getting buy-in about what it is you’re trying to figure out around these tools.
The real hard part is that we’re in the middle. AI has shown up, it has gotten ubiquitous very quickly, and it’s continually changing. We’re in this space between when there were clear standards and when there will be clearer standards around AI. But I think we’re going to be in this middle space for a while. It’s going to take us much longer to figure it out.
If we look at the internet as an example, it took us really 15 to 20 years to figure out what the some of the best practices, appropriate practices, of a variety of different things are as it relates to scholarship and writing. Though, to be fair, I would say that’s still a work in progress. We are going see some of the lines continue to change and be comfortable with that.
That instability is the reseason I emphasize transparency about our use until norms are more clearly established. Until we know otherwise, transparency helps us see what is actually happening and how these tools are being used. It also helps us understand where the lines are, because if I’m transparent, and I did something, and it doesn’t quite work, well, that’s a great conversation. There’s great opportunity for learning. It’s going to be a little bit uncomfortable and it may cause some trouble if I did so without due consideration. But knowing we might have crossed the boundary, is valuable because we’re all trying to figure it out. You’re better off being transparent right now, so you can better understand where the lines are.
Who are you?
At the end of the day, all of this boils down to: what kind of scholar, researcher, educator, scientist, writer, etc do you want to be? No matter what is covered today, all the different tips, tactics, and guidance that is out there — none of that matters unless you can answer this question and hold yourself to that, or find a way of having others help hold you to the person that you want to be, the person you’re aspiring to be.
That’s not easy, but more than anything else, that’s the true work; that’s the practice. The temptation is real and for many of us, we’ll need to develop those integrity muscles around this particular aspect. And there may be mistakes, but you, at the end of the day, know that you are being true to the person, to the scholar, to the researcher that you really see yourself as.
Mapping and defining your work
These tools have some great possibilities but we have to be careful that the skills we’ve developed do not atrophy. Many of our skills are like our muscles: if we don’t use it, we lose it. We sometimes think of skills like riding a bike; once I learn it, I never forget it. I don’t know if that’s true, but if it is, there’s still the matter of distance. If I haven’t ridden a bike in 10 years, I can probably still ride it, but how far I can go will vary. In fact, it will vary in accordance with whether I’ve been doing any other exercise (.i.e. adjacent skill development).
Recognizing the skills impact, it means we should raise our reflective attention to our usage. Ask yourself with each use: Why am I using this? What’s leading me to use this? What’s the value behind that, and what’s the scale?
This isn’t easy, it’s a practice we’re trying to build. It’s cognitively demanding to keep challenging and questioning ourselves. I run into that all the time. Interrogating ourselves is hard. But it can be supported. Take an hour and map out:
What is my intention in being a research, scholar, writer?
What do I want to provide or achieve in my work?
What is your line?
Why?
How will you know that you’ve crossed it?
What is yours to hold onto?
What do you feel comfortable exporting to AI?
What are 5 ways (from easiest to hardest) that you can hold yourself accountable to your line?
Through that exercise, consider mapping it out—even in a flowchart or checklist that serves as a shorthand for how you want to show up and what you want to do. It might include things like “I’m going to use AI for these reasons and at these points, and I’m going to use myself at these points.”
For example, I will use AI to surface my own thinking. I’ve set up my AIs to interview me and to challenge my thinking. It’s not creating the content or the ideas; it’s getting me to explore and develop connections through asking me questions. What I’m not asking AI to do is to create the content of my talks and writing.
Some straightforward guidance?
I think I struggle with a lot of practical guidance on this because it really does start with you first deciding where’s your line.
The most straightforward tip is to look at where you want to publish something. If you know you are going to write something or submit research, one practical tip—even before AI—was to determine which publication you want to write for. Given the different styles, expectations, etc, it can save you time and energy as you both conduct your research and prepare your submission. And many journals now have statements about what level of AI use is appropriate. If they don’t, it can be as simple to email the editor for clarity.
Be prepared to see that a lot of journals are going to restrict AI use to copyediting uses. That makes a certain sense. Academic publishers are conservative by nature; they tend to preserve tradition, playing the prestige game, and making profits. Because AI practices are still new and unstable, and AI content raises questions about copyright, most publishers are being deliberate.
One way to think about this—and it’s unfortunate—is to think defensively about how you can prove your work. If challenged, what is the paper-trail of effort you can point to. One of the easiest ways is working in documents that have version history. I’ve been myself working in Google Docs for, I don’t know, close to 20 years now. So I do think there is something within this of just thinking about: where do I collect my thoughts, and how, if asked, can I show my work. I think that’s probably going to be the easiest at this moment: just using a tool that actually does version history. I might also include downloading or screenshots with timestamps of my searching history related to research. As I mentioned, including chatlogs can also help with better clarifying what you did or didn’t do.
That’s not a lot of big advice, but it’s start. If you have practical advice, please share it in the chat!
Finally, in writing this piece, it made me realize that I have some strong ideas about doing a workshop about how to surface and identify your line. If that sounds appealing, please complete the poll below to let me know there’s actual interest to run it.
The Update Space
Upcoming Sightings & Shenanigans
Keynote speaker at the Reimagining the Liberal Arts in the Age of AI Conference, July 21-23 at the University of Mary Washington.
EDUCAUSE Online Program: Teaching with AI. Virtual. Facilitating sessions: ongoing
Recent Recordings, Resources, & Writings:
Damm, C., & Eaton, L. (2026, March). From prompt to practice: A framework for transparent GenAI use in higher education. EDUCAUSE Review.
Eaton, L., Nemeroff, A., & Sun, X. (2026). AI-assisted course design and development. In K. S. Ives, M. Cini, & R. Schroeder (Eds.), AI applications in online higher education administration: Strategies for maximizing returns and improving outcomes. Routledge.
Margin of Thought with Priten: Season 1, Episode 5: How Can We Center Pedagogy During the AI Tech Wave? (February 2026)
Online Learning in the Second Half with John Nash and Jason Johnston: EP 39 - The Higher Ed AI Solution: Good Pedagogy (January 2026)
The Peer Review Podcast with Sarah Bunin Benor and Mira Sucharov: Authentic Assessment: Co-Creating AI Policies with Students (December 2025)
David Bachman interviewed me on his Substack, Entropy Bonus (November 2025)
The AI Diatribe Podcast with Jason Low (November): Episode 17: Can Universities Keep Pace With AI?
The Opposite of Cheating Podcast with Dr. Tricia Bertram Gallant (October 2025): Season 2, Episode 31.
The Learning Stack Podcast with Thomas Thompson (August 2025). “(i)nnovations, AI, Pirates, and Access”.
Intentional Teaching Podcast with Derek Bruff (August 2025). Episode 73: Study Hall with Lance Eaton, Michelle D. Miller, and David Nelson.
Dissertation: Elbow Patches To Eye Patches: A Phenomenographic Study Of Scholarly Practices, Research Literature Access, And Academic Piracy
AI Syllabi Policy Repository: 200+ policies (always looking for more- submit your AI syllabus policy here)
Finally, if you are doing interesting things with AI in the teaching and learning space, particularly for higher education, consider being interviewed for this Substack or even contributing. Complete this form, and I’ll get back to you soon!
We periodically host small-group workshops and leadership sessions for higher ed teams. You can learn more about our current offerings here.
AI+Edu=Simplified by Lance Eaton is licensed under Attribution-ShareAlike 4.0 International
