Dear students,
It’s the start of another quarter, and again, I am facing what has become a familiar and constant dilemma. How do I address AI?
Okay, here’s one idea—what if I paint a nightmare scenario for you? It goes something like this—in an imagined near future, I use AI to generate an assignment, you use AI to write an essay for my AI-generated assignment, and then I use AI to generate feedback on your AI-generated essay. Pretty soon, the class is completely taught by and attended by robots without any need for human teachers and students, and before we know it, we’re living in a dystopian world ruled by robot overlords.
But this piece is not titled “Your Teacher’s Esotoric Fears of a Bleak, Nihilistic Futurescape.” It’s titled “Why AI Is Bad for Your Grade,” and one big reason is that I want it to feel relevant to you. I am fully aware that a lofty and abstract piece like why AI is bad for your creativity or brain or humanity might not grab your attention. But, to be clear, I do believe that AI is problematic for all those reasons.
And also, AI is bad for your grade.
When I say that AI is bad for your grade, I do not mean that I plan to give you a punitively low grade for AI-generated writing. I simply do not have the time and energy to try to police every assignment that I suspect has been written by a robot. And also, I don’t think punishment is effective, responsive, or humane (I talk more about these beliefs in my book Anti-Oppressive Universal Design for Teachers: Building Equitable Classrooms). In fact, I have stopped giving a “0” on essays or reporting students to administrators.
So does that mean there are no negative effects for using AI? Um…no. Let me explain. Although I do not penalize students for using AI, there are still natural consequences for using AI that will lower your grade. Now, I get it if you are scratching your head and saying, “Huh? What does that mean?” Punishment and consequence are often conflated, so let me try to explain the difference through two hypothetical situations.
In Scenario A (punishment), I read a student essay that I believe to be AI-generated. Perhaps I even use an AI-checker to support this assessment. Then, I give the student a “0” on the paper, which is likely to result in failing the class. In addition, I report the student to school administrators, which may result in academic probation. In other words, I am passing judgement on the student’s honesty and calling their ethics or morals into question. The result of this is that the student feels shame and is unlikely to grow or learn from the experience. That is punishment (And in case it’s not clear by now, I do not believe in judgement and punishment as effective teaching tools).
In Scenario B (consequence), I read a student essay that has fake quotes attributed to a source that never actually said or wrote those quotes. The essay also cites sources that don’t seem to exist. In addition, the ideas expressed in the essay are general and vague. The content of the essay also seems disconnected to the outcomes of the assignment and the learning of the class. There is also a summary of a scene in a novel that never happened with characters that don’t exist in the novel. In other words, the essay contains common AI hallucinations and fake content. Although I might suspect that the essay is AI-generated, I am not grading the student to punish the student for having AI content. I am grading the student on the content of their essay. However, the fact that the content of the essay does not demonstrate the outcomes of the assignment, contains fake support, and is vague and general—is a consequence of using AI.
Sadly, I have had students come to me after submitting an essay with fake sources or a summary of scenes and characters that never actually appear in the story they are analyzing and demand to know, “Why did I only get a 65/100 when there are no grammar errors in the essay?”
My heart always aches for these students because they have been told that the only thing of value is that their writing be grammatically “correct” and that this is more important than the expression of their own voices. With damaging messages like this, it is no wonder that some students resort to AI.
In fact, I do not care about grammatical “correctness” (which is a restrictive social construction normed on white-centric standards). I do care about writing that has meaning to you. I want you to write in a way that allows you to express your full humanity and unique voice. Writing should allow you to be creative, get messy, think through ideas, and have fun. From my viewpoint, it is impossible for bland, generic grammatically “correct” AI-generated writing to be more important than your own creative and meaningful expression of ideas.
So, I realize that I still haven’t fully answered my hypothetical student’s question—why did this grammatically “perfect” AI-generated essay get a 63/100? After all, I claim that I am not grading punitively on suspicion of AI usage, so how did the robot get such a low grade?
I’ve alluded to common pitfalls of AI like hallucinations resulting in fake quotes, and I think a brief explanation of my grading criteria can answer this question in even more detail and also answer the larger question of this piece—Why is AI bad for your grade?
Some details about how I grade:
- I give everyone full points in the category of “Writing clarity.” That means an essay with the glossy sheen of AI “error-free” writing and the essay with human writing choices, which may include run-ons, comma splices, and dangling modifiers—get the exact same full credit in this category. In other words, AI-generated grammar is not an advantage in fulfilling this criteria.
- I also give everyone full points in a category I have designed called “Voice and Vision.” This is simply an acknowledgement that each writer’s own voice and vision is important and deserves to be celebrated and valued. Actually, I’m not even sure what an AI “voice and vision” is, but using AI certainly does not give you an advantage in fulfilling this criteria.
- The most substantial criteria that is worth the most points is “Ideas and Content,” and here is where AI-generated content truly shines at being…mid. AI is like an indiscriminating scavenger. It steals from what is out there in cyberspace and isn’t picky about what it takes and spits back out in the form of an essay. These regurgitated ideas might sound okay, but they don’t really say anything. And if multiple students are putting similar prompts in the chatbot, then the same phrases and sentences are going to turn up over and over again so the person reading the essays (me) has the uncanny experience of reading vaguely similar essays that contain nothing really meaningful or unique.
- The next most substantial criteria in my grading rubric is “Support,” and the false content resulting from AI hallucinations often means low points in the category of “Support” as well as “Ideas and Content.” If you’re curious about an example of an AI hallucination, I’m happy to give you one! To test out just how accurate (or inaccurate) AI can be in generating information, I decided to ask an AI chatbot about myself. The prompt I used was something along the lines of “Who is the author Diana Ma?” Even though I knew AI would get some things wrong, I have to admit that I was shocked by what AI generated. According to AI, I was apparently a writer for both Desperate Housewives and America’s Next Top Model! (I am sad to report that this is false).
I want to make one more point about AI hallucinations that doesn’t really have anything to do with grading but just has to be said. Although the fake information about me being a reality TV writer is amusing, less amusing is what AI generated about my debut novel Heiress Apparently. I wrote this book to resist flat, one-dimensional stereotypical representations of Asians, and guess what AI did to the plot, themes, and characters of my book? If you guessed that AI reduced the plot, themes, and characters to flat, one-dimensional stereotypical caricatures—then you would be right.
And now back to details about how I grade:
- I offer flexible options in the assignment so that you are engaged by the writing you’re doing! If there’s something that’s not working for you in the assignment, come talk to me, and we can make it work. I want the writing you do to be meaningful for you, so I’m more than happy to make adjustments.
- I am happy to offer extensions and support! I understand that some students turn to AI because they are confused about an assignment or are not confident in their writing skills. Please know that I am here to help and would much rather give you an extension and answer questions about an assignment than give a low grade on an AI-generated essay.
Trust me when I say I derive no joy in giving low grades. In fact, I hate giving students low grades, so I am personally invested in encouraging you to submit human-written work so I can give high grades.
Actually, that brings me to an important point. My students who turned in essays with tangents, sentence fragments, punctuation “errors,” risky claims, surprising conclusions, unique wording, and other human writing—were doing great, both in their writing and in their grades. When you have a sense of agency and pride in your work, your writing feels fresh and meaningful and is simply a joy to read—regardless of any “imperfections.”
Take this piece of writing. I’m putting up this blog post up on my website pretty quickly since I want it to be available to students at the start of the quarter. That means that I didn’t have the time to edit carefully or ask others to edit. There are probably proofreading errors and definitely wording that can be clearer, and maybe even some underdeveloped ideas and extraneous sentences or tangents. In other words, this is not a perfect piece of writing. I doubt that any of my writing is perfect–whatever that means. But I didn’t write this piece to produce perfection.
What matters is that this blog post is mine, that it does what I want it to do, and that it means something to me. That’s what I want your writing in this class to be for you.
I wrote this blog post because I needed something that would explain the effects of AI on your grade and I thought it would be interesting and useful to write it myself. After all, the act of writing this blog post gives me the chance to think through ideas about the teacher I want to be, and it allows me to have fun in the creative process. Yes, I could have used AI to generate something about why AI is bad for your grade, but other than the fact that I’m opposed to pass off content stolen and compiled by AI as my own writing–I also feel that using AI wouldn’t result in a very interesting blog post. It probably would be less fun to read (and no fun to write since I wouldn’t be writing it). It would certainly be less specifically relevant. In short, I hope that you are enjoying and learning from the blog post, and I believe I have increased the chances of both by not using AI.
Okay, I want to come back to my hypothetical student who got a 65/100 on their essay. To be honest, when I was first dealing with the deluge of AI-generated essays, I was overwhelmed. Back then, I offered rewrites, but so many students were getting low grades (like a 65/100) due to AI (for the reasons outlined above) that I would get a large number of rewrites that were again generated by AI. It was so demoralizing and such a drain on my time and energy that I stopped offering rewrites.
Logically then, it follows that writing generated by AI, which does not give you agency or a sense of pride, also does not produce fresh or meaningful writing and certainly is not enjoyable to read. It also stands to reason that I would not be eager to open the door to AI-generated rewrites. But that left me with a problem.
Consequence might not be the same as punishment, but without a means for repair, consequence can sure feel like punishment. In other words, the student who got a 65/100 needs a means for repair. That is why I have begun to offer rewrites again, but with one important caveat—that this is an opportunity for true repair and not just an opportunity to fix your grade with a slightly less problematic AI-generated rewrite.
So, here’s what we’re going to do. If you get a low grade and want to rewrite your essay, we will first have an honest conversation free from recrimination or judgement where I don’t lecture you about honesty or ethics (and maybe you don’t leave me AI-generated negative “Rate My Professor” reviews because you got a low grade on AI-generated writing). Cool? Great!
This is how our conversation will go…
We might start by discussing what kind of barriers to your writing might have made you resort to AI (again, without shame or judgement). Then we can collaboratively figure out how to remove those barriers. I envision this as a supportive dialogue that focuses on helping you value your own voice and ideas. It is my hope that after this conversation, you will be able rewrite your essay with a sense of engagement and ownership of your own writing, free from the robot’s control (Sorry, I’ve slipped back into dystopian metaphors again).
But speaking of dystopian metaphors, the topic of this blog post is partly about finding joy in our own writing, so I’m going to wrap up by indulging in dystopian metaphor simply because it’s fun. (Yes, I realize that this definition of “fun” is purely subjective)!
Okay, ready? Here we go.
Your voice and your writing matter, so don’t let the robots steal that from you. Don’t let the soulless language of machines replace your humanity and creativity. It’s us against the robots, and I’d like to think we can win.