Lately, I’ve noticed something interesting in my accounting courses: students are increasingly using AI to complete their take-home assignments. That, in itself, doesn’t surprise me—in fact, I think it would be unwise not to use AI at this point. The technology is here, it’s powerful, and it’s quickly becoming a standard tool in the professional world.
What has been interesting, though, is the reaction when I take points off.
One student recently challenged me directly. She realized that I had begun using AI as part of my grading process—reviewing her answers with it, comparing reasoning, and using it as a second set of eyes. Her argument was simple: it was unfair for me to use AI to grade her work.
I’ll admit, I found that perspective interesting.
My response was straightforward: I don’t see it as unfair at all. I review what the AI suggests, I evaluate it, and I make the final decision. In other words, I’m doing exactly what I expect my students to do.
The real issue wasn’t the use of AI—it was that her score was lower than expected.
And here’s where things get a bit ironic.
I’m confident that most, if not all, of my students are using AI to help complete these assignments. I don’t discourage it—in fact, I encourage it. Students—and professionals—who ignore AI risk falling behind. The world they are entering will demand fluency with these tools.
But there’s a critical misunderstanding happening.
Many students assume that the first answer AI gives them is good enough.
It usually isn’t.
AI is a starting point, not a finished product. The first output is often generic, incomplete, or slightly off-target. What separates strong work from mediocre work is what happens next.
This is where critical thinking comes in.
Students need to evaluate the AI’s response:
- Is it actually answering the question?
- Is it precise and technically correct?
- What’s missing?
- How can it be improved?
From there, they refine their prompts, ask better questions, and iterate. This process—often called prompt engineering—is a real, learnable skill. And like any skill, it improves with practice.
This isn’t just a classroom issue. In professional settings—especially in work that requires judgment, analysis, and defensible conclusions—the first answer is rarely the final answer. Whether reviewing financial information or forming an opinion that others will rely on, the ability to question, refine, and validate information is what ultimately creates value.
That’s the point I want my students—and really anyone using AI—to understand:
AI alone does not produce high-quality work. AI combined with critical thinking does.
As educators, this creates a challenge—but also an opportunity. Instead of trying to eliminate AI from the learning process (which is unrealistic), we need to design experiences that require students to think, evaluate, and refine.
Because in the end, the goal isn’t just to get the right answer.
It’s to understand why it’s right—and how to improve it when it isn’t.