Assessing Student Learning in the Age of AI: 3 Ways to Make Learning More Visible 

| Blog Articles | Share:  

As AI changes what faculty can infer from finished assignments, higher education institutions are reconsidering what evidence best demonstrates student learning. 

AI can help students produce work that is polished, persuasive, and technically correct. But the quality of a student’s essays, projects, or presentations does not always reveal how well they understand the material—or whether they can apply that knowledge independently. 

That makes this bigger than academic integrity. It goes to the heart of assessment: What evidence demonstrates that learning has occurred, and how confidently can it be attributed to the student? The answer will vary by course, discipline, and stakes, but the principle holds. Institutions need evidence of reasoning, decisions, and application, not just the finished product. 

Here are three ways institutions can assess student learning and gather meaningful evidence of what students know and can do.

1. Ask Students to Explain the Thinking Behind Their Work 

A student’s finished work shows what they submitted. An explanation can reveal how they arrived there. 

Short oral defenses, follow-up questions, and recorded explanations give students an opportunity to articulate their reasoning in their own words. Faculty might ask: 

  • Why did you choose this approach?  
  • What alternatives did you consider?  
  • What information most influenced your decision? 
  • What would you change if one of the underlying conditions changed? 

These moments do not have to become full-scale oral examinations. A few well-chosen questions can reveal what the submitted work alone cannot. Consider a nursing student who selects the appropriate intervention in a clinical scenario. The answer matters, but so does the student’s ability to explain which symptoms informed the decision, what risks they considered, and what would cause them to change course. That explanation makes clinical reasoning more visible. 

This approach distinguishes between a student who can recognize a plausible answer and one who understands why it is appropriate. 

2. Create Opportunities to Apply Knowledge Under Changing Conditions 

Many assignments ask students to produce an answer based on a fixed set of information. In practice, however, circumstances rarely remain fixed. 

Applied assessments can introduce new information, changing constraints, or competing priorities that require students to adjust. Students must then demonstrate how well they can apply what they know. 

An engineering student, for example, might develop a technically sound solution to a design problem. Rather than ending the assessment there, faculty could introduce a change that affects the student’s original solution. A material might become unavailable, or the operating conditions might change. The student would then need to determine whether the original design still works, make any necessary revisions, and explain the reasoning behind those decisions. 

AI may still play an appropriate role in assessing student learning. The adaptive step is where faculty see what the student can do once the situation no longer matches the prompt. 

The ability to adapt matters most when students need to exercise judgment rather than arrive at a single correct answer. 

3. Capture Evidence of Learning Along the Way 

Evaluating only the finished product can leave faculty with little insight into the decisions and reasoning that shaped it. Formative assessment can create greater visibility by giving faculty opportunities to gather evidence as learning develops. Institutions might identify meaningful checkpoints as students work through an assessment: an initial proposal, a research plan, a draft, a decision log, or a reflection on how feedback changed the work. 

 The goal is not to collect every keystroke or turn each assignment into a surveillance exercise. It is to capture the moments that best demonstrate the student’s intellectual contribution. 

A capstone project offers a useful example. In addition to evaluating the final presentation, faculty might ask the student to defend a key decision, explain how the evidence shaped the recommendation, and identify what they would do differently with more time or information. 

Where Should Institutions Begin? 

Assessing student learning in the age of AI does not require redesigning every assessment at once. A more practical starting point is to focus on the places where the consequences—and the risk of drawing the wrong conclusion—are greatest. 

Assessment leaders and faculty can begin by asking: 

  • Where could a student use AI to produce a credible final answer without demonstrating the intended learning? 
  • Which assessments carry the greatest consequences for progression, graduation, professional preparation, or public trust? 
  • What would faculty need to observe to evaluate the student’s reasoning, judgment, or ability to apply knowledge? 
  • Which uses of AI are appropriate for the task, and are those expectations clear to students and evaluators? 

The right response will not be the same for a first-year writing assignment, a nursing simulation, and a final capstone. Each makes a different claim about what the student knows or can do. The evidence should reflect that claim. 

Every credential makes a promise. The assessments behind it should provide the evidence needed to support that promise. 

Build Greater Confidence in Assessment Results 

Meazure Learning helps higher education institutions build assessments that produce credible evidence. Contact us to discover how.