How to Prepare for the New University Assessments: Explain Your Thinking, Not Just Your Answers

Imagine submitting an essay you spent a week writing, only to be asked one simple question: “Why did you make this argument?”

Could you explain your answer without opening the document?

For a growing number of students, this isn’t hypothetical. Across higher education, some instructors are rethinking how they assess student learning in response to generative AI—adding oral defenses, staged submissions, and in-class writing to the more familiar essay or final exam.

The change isn’t universal, and it isn’t happening at the same pace everywhere. But it’s real enough to plan for. And it asks something different of your studying: not just can you produce the work, but can you explain how you got there.

Why Some Instructors Are Rethinking Assessment

For decades, assessment rested on a quiet inference: a good essay, a correct problem set, a polished report was treated as evidence of good thinking. The product stood in for the process.

That inference has become less reliable.

A 2026 Nature survey of 1,054 UK undergraduates found that roughly 94% use generative AI to help with assessed work, and 12% directly insert AI-generated text into their coursework. A separate Nature feature, published in September 2026, describes how instructors are responding—with oral examinations, in-person presentations, in-class writing, and review of revision histories.

The issue isn’t simply cheating. It’s structural: when a polished response can be produced on demand, the artifact carries less information about whose reasoning produced it. So some instructors are collecting different evidence.

A note on scope: this is a response at the level of individual courses and departments, not a coordinated national policy. Many courses still assess final products exactly as they always have. But if even one of your courses uses an oral defense or a staged submission, it’s worth knowing how to prepare for it.

What the New Assessment Can Look Like

The shift, where it happens, is broadly from assessing only the final product to also assessing the process that produced it. It takes several forms.

Oral and conversational exams. Students explain their reasoning live, often while working through a problem. A 2026 paper by Lorena Barba and Lecia Stegner describes a “conversational exam” format designed to make oral assessment workable at typical class sizes—tested with 58 students in small groups over two days—as a way of hearing students reason in real time rather than inferring it from a submitted file.

Staged submissions. Instead of one final file, you submit drafts, notes, and revision history, so the trajectory of your thinking stays visible.

Reflective justification. You’re asked to explain why you made the choices you did—methodological, interpretive, or structural.

AI-aware assignments. Some instructors build AI into the task itself: asking students to use several models, compare their outputs, and critique where they go wrong.

The common thread: assessment is becoming more than a judgment of the final product. The final product still matters—it’s just no longer the only thing being read.

What This Means for How You Study

If assessment asks you to explain and defend your reasoning, then studying that only prepares you to produce an artifact is incomplete.

Here’s what changes.

Shift 1: From “I can produce it” to “I can explain it”

The old test was whether your final answer was right. The additional test is whether you can walk someone through how you got there—including the dead ends, the choices you rejected, and the assumptions you made.

A practical way to check: After you finish a problem or a draft, close everything and explain your reasoning out loud, as if to a classmate. If you can’t narrate the why behind each step, you’ve found the gap an oral exam would find.

Shift 2: From “final product” to “visible process”

If you’re assessed on staged submissions and revision history, then keeping a record of your thinking becomes part of the work.

That doesn’t mean saving every scrap. It means keeping the versions that show how your argument developed: your first outline, the draft where you changed direction, the notes where you worked out a difficult concept.

A simple habit: At the end of each study session, write two lines—what you understood, and what you’re still unsure about. Over a term, that log becomes a map of your learning that you can actually talk about.

Shift 3: From “AI-assisted answer” to “verifiable understanding”

This is the most important shift, and the least comfortable.

The educators quoted in the 2026 Nature article we mentioned earlier weren’t trying to ban AI. They were trying to design assessments where your reasoning is what shows up—not the tool’s output. That’s why so many of the examples involve open problems with multiple valid approaches, where the student has to justify a choice rather than arrive at a single correct answer.

The implication: Using AI to produce work is not, by itself, a strategy. Using AI to build understanding you can then demonstrate is.

The 5 Questions You Should Be Able to Answer About Your Own Work

This is the most practical part of preparing for a process-based assessment—and it works for an essay, a project, a lab report, or a problem set.

Before you submit anything, try to answer these five questions without notes:

1. What is the main argument or solution?
State it in one sentence. If you can’t, you don’t have one yet.

2. Why did you choose this approach?
Not “why is it correct,” but “why this one, over the alternatives.” This is the question oral examiners ask most often.

3. What alternatives did you consider and reject?
Being able to name the path you didn’t take shows you understood the choice, rather than stumbled into it.

4. What evidence supports your conclusion?
Point to the specific source, data point, or derivation—not the general topic.

5. What would change your answer?
This is the hardest one. It tests whether your conclusion is a considered position or a fixed belief.

If you can answer all five, you’re ready for a conversation about your work. If you can answer two or three, you now know exactly what to rehearse.

How to Prepare: A Practical Sequence

Here’s a sequence you can apply to any course where the final assessment includes explanation, oral defence, or staged submission.

Step 1: Study for retrieval, not recognition.
Close the book and write down what you remember. Answer questions without notes. The gap between what you recognize and what you can produce is exactly what an oral exam will expose.

Step 2: Practice explaining aloud, before you have to.
Explain a concept to a friend, a recording app, or an empty room. If your explanation stalls, you’ve located a gap. Repeat until it flows without notes.

Step 3: Keep a learning log.
Two lines per session—what clicked, what didn’t. This gives you both a revision tool and a record you can draw on if you’re asked to discuss your process.

Step 4: Rehearse the conversation, not just the content.
Oral assessments aren’t only about knowing the material; they’re about navigating questions about it. Ask a friend to interrupt with “why?” and “what if you’d chosen the other approach?” Practice responding without defensiveness.

Step 5: Use AI as a rehearsal partner—for questions, not answers.
Instead of asking AI to explain the topic, ask it to ask you questions about it. That turns a tool that can produce answers into a tool that prepares you to give them.

Where StudyWizardry Fits In

The shift toward assessing process rewards studying that makes your understanding visible—to yourself, first.

Study Sets let you turn your own materials—lecture slides, readings, notes—into a study guide, lesson outline, flashcards, and quizzes. That’s useful here because it gives your material a structure you can narrate. When you’re asked to explain your reasoning, having a clear structure to draw on is half the work.

Quiz Generator turns your notes into practice questions. That’s retrieval practice, and it’s also a rehearsal for the moment someone asks you to explain. Answering without notes builds the kind of understanding an oral exam is designed to detect.

Flashcards help you keep foundational concepts and definitions fresh, so that when you sit down to reason through a problem or prepare an explanation, the basics are already available to you.

Study Planner helps you spread retrieval and explanation practice across the term, rather than cramming it into the week before an oral defense.

And because StudyWizardry works in your browser and syncs across web, iOS, and Android, you can rehearse an explanation, check a concept, and return to your notes without losing your place.

The principle is consistent: the tool supports the thinking. The explanation is yours.

The Honest Truth

Assessment is changing in some places because the old signals became less reliable—and that’s uncomfortable for students and instructors alike. One computer scientist told Nature plainly: “I don’t think anybody has a ‘right’ answer right now.”

But the direction is clear enough to prepare for. Where assessments ask you to explain, defend, and reconstruct, the students who adapt fastest won’t be the ones who use AI least. They’ll be the ones who use it to build understanding they can demonstrate without it.

Your next study session, try this: After you finish a chapter or a problem set, close everything and answer the five questions above out loud. Then note where you hesitated. That hesitation is what a process-based assessment would find—and what you now know to work on.

📚 References

• From Trojan horses to AI-proof exams: how professors are tackling students’ AI use. (2026, September 1). Nature. https://www.nature.com/articles/d41586-026-02370-2

• Barba, L. A., & Stegner, L. (2026). The Conversational Exam: A Scalable Assessment Design for the AI Era. Computer, 59(6), 132–139. https://arxiv.org/abs/2601.10691

• Process-Oriented Assessment. (2026). AIED Knowledge Base. https://edtechdev.github.io/aied/concepts/process-oriented-assessment/