Start Where It Hurts: Using AI to Tackle Assessment Pain Points
How AI Makes Assessments Less Painful and More Purposeful
AI can help treat assessment pain points with creativity, differentiation, and clarity. Start where it hurts, and let the healing begin.
ChatGPT Prompt: Create a square-format illustration in the style of a muted-color graphic novel. The scene takes place in a doctor’s office. A female teacher sits anxiously on an exam chair with her hands folded, looking slightly overwhelmed. Seated across from her is a friendly doctor in a white coat with a clipboard and stethoscope, speaking with care and focus. Above the doctor’s head are three speech bubbles that read: “Pain Points,” “Differentiation,” and “Clarity.” On the wall behind the teacher is a framed cartoon-style poster that says, “Where does it hurt?” and shows a person pointing to their head. The overall tone should be warm, slightly humorous, and professional—like a metaphorical consultation for improving assessments.
Dani Kachorsky, PhD
Recently, at Phoenix Fan Fusion, I had the chance to sit on a panel about AI in education. During the Q&A, a teacher raised a concern I’ve heard so many times before. He said, “There are just so many tools out there—I don’t even know where to start.”
Totally fair. If you’re even remotely paying attention, it feels like every other day a new AI tool pops up promising to revolutionize your workflow, save you hours, and brew your morning coffee. Overwhelming doesn’t begin to cover it.
So here’s what I told him—and what I’ll tell you: Start with your pain point. That’s where AI is going to feel most valuable to you. For me? My pain point has always been traditional assessment.
Why Assessment Is My Pain Point
Let me be clear: I love planning lessons. I love designing learning experiences, building in discussion and collaboration, and watching students engage. But assessments? Traditional ones? They’ve never been my thing.
Multiple choice tests feel like memorization drills—and in the real world, memorization is rarely required. Five-paragraph essays? That format doesn’t exist beyond school walls. And don’t even get me started on grades—reducing students to letters and numbers just feels wrong.
But sometimes, we have to create traditional assessments. School or district mandates, standardization, accountability—you know the deal. And that’s exactly where AI has helped me the most.
How AI Can Create Traditional Assessments
This is where large language models like ChatGPT and Claude shine. Traditional assessments are a perfect match for large language models. Why? Because most of our go-to assessments—whether we’re talking about multiple choice tests, short answer questions, essay prompts, or discussion questions—they’re all built around language. And language is what these models are trained on.
Here’s how I have used AI to create traditional assessments and how you can too:
Discussion Questions: Ask your AI to generate questions about a topic, text, article, or video. Want to ensure accuracy? Feed it a PDF, transcript, or excerpt. Direct the AI to stick to what you provided, so you’re not getting questions about parts of the material you haven’t taught yet.
Multiple Choice / Short Answer / True-False: Give the AI the source material. Then, give it parameters. How many questions do you want? How many answer choices? What types of questions? Do you want an answer key? Boom. You’ve got a draft of a test.
Essay Prompts: Ask the AI to create a range of prompts on a specific topic or text. If you want the prompts in a particular standardized structure, such as the AP Language & Composition prompts, feed the AI sample prompts along with your prompt draft, and it will shape yours to match the structure.
Rubrics: Upload your assignment and instruct AI to draft a rubric. You can list the categories or ask it to suggest categories based on the assignment description. This can be eye-opening because the AI might generate categories you did not expect. This might mean the assignment isn’t as clearly focused on a particular skill or standard as you originally thought.
If you would like to see one teacher’s process for creating traditional assessments using some of the suggestions above, check out our last post in which my husband, Jeremy Kachorsky, shares his workflow with and advice for using AI to create “just right” tests for his Biology and Chemistry classes.
How AI Can Revise Assessments
What really surprised me was how helpful AI is after I’ve written an assessment. There’s something magical about feeding it an old quiz or essay prompt and asking it to update the language, or integrate vocabulary we’ve added to the curriculum. I’ve had AI reword test questions so they’re clearer, refresh rubrics that felt clunky, or restructure essay prompts to better align with our standards.
Say you’ve got an old assessment that could use a refresh. AI can:
Update vocabulary to match what you now emphasize in class.
Reformat for clarity or consistency (especially helpful if different teachers drafted materials).
Rewrite prompts or questions to match updated standards or test styles.
Tweak rubrics—adjust point values, reword criteria, add categories you missed.
One fun approach? Ask AI to make revisions, but instruct it to ask clarifying questions first. This tip, courtesy of my brilliant colleague Shelby Stringer, helps you refine what you actually want students to show you. The AI might ask, “What weighting do you want for each rubric category?” or “Do you want this test to be all multiple choice, or should I mix in short answers?” or “How much focus should be placed on grammar, spelling, and punctuation?”
I love this because it forces me to think through details I might have overlooked. If the AI is confused, there’s a good chance my students would be too. And the back-and-forth makes the final product stronger.
How AI Can Differentiate Assessments
This is the part where I think AI really shines. We all know no two classrooms are the same. You might have students working at grade level, alongside English language learners, alongside students with IEPs or 504 accommodations. Differentiation shouldn’t be considered a luxury; it’s just what good teaching looks like, but it can be a challenge for overworked teachers.
This is where AI has saved me hours. Once I’ve created a core assessment, I can ask AI to adjust it in a variety of ways depending upon my students’ needs. Some possibilities include:
Lowering the reading level for ELLs or students reading below grade level.
Reducing the number of questions for students with accommodations requiring fewer items.
Simplifying the language without watering down the concepts.
Shifting question formats—for example, turning open-ended questions into matching or true/false.
And I don’t have to start over each time. I feed the AI my original test, tell it the change I need, and it produces a draft I can tweak. The more context I provide (such as examples of what a sixth-grade reading level entails or a list of topics that must be covered), the better the results.
AI doesn’t replace my judgment here—it just helps me do the work faster so I can spend more time fine-tuning the details that matter.
So, How Do You Start?
If this sounds useful, but you’re not sure where to begin, here’s my advice:
Pick your biggest pain point. For me, it was assessment. For you, it might be something else—lesson planning, rubrics, discussion guides.
Choose one tool to experiment with. Don’t get caught up in the sea of options. ChatGPT, Claude, Copilot—pick one and dive in.
Give it source material whenever possible. PDFs, transcripts, articles—it helps AI stay focused on your content.
Be clear in your prompt. Tell it what you want. And better yet? Tell it to ask you clarifying questions before it generates anything.
Always review before sharing with students. AI saves you time creating, so you have time to double-check for accuracy and make tweaks.
Just remember: AI isn’t going to do our jobs for us. But it can absolutely make parts of the job easier, especially the parts that sap our energy. For me, that’s assessment. Maybe for you it’s something else. But start where it hurts, and let AI help lighten the load.
AI Transparency Statement
This post was created using the AI-assisted workflow I described in a previous essay. I began by audio recording my thoughts and experiences, then used AI to transcribe and synthesize my reflections while maintaining my voice. I added and revised material through a few additional prompts in the LLM interface before copying the content into a document, where I made further revisions. I read the sources that are linked and cited in this post in advance of my audio recording and verified the accuracy of information prior to posting the essay.
AI Prompts
ChatGPT prompt for blog post: Below is a transcript of an audio recording in which I discuss how AI can be used to create traditional assessments (e.g., multiple choice tests, essay prompts, discussion questions, etc.) and how AI can be used to differentiate assessments based on students' needs (e.g., disability accommodations, reading level, English language learning, etc.). Create a Substack/blog post using the material from this transcript. Begin with an introductory anecdote from Phoenix Fan Fusion that shares my recommendation to begin using AI for a teaching pain point. For me, this is often traditional assessments because I don't find them to be reflective of real-world practices, and therefore, I don't personally find them valuable. From there, have a section on how AI can be used to create assessments, a section on how AI can be used to revise assessments, and a section on how AI can be used to differentiate assessments. End with a getting-started section with bullet-pointed or numbered how-to advice. Utilize my informal Substack tone. You can reference our previous chats about Substacks and blog posts to replicate this tone.