A student completes a project, submits it, and receives a grade. The work looks polished. But how much of it can the student explain, adapt, or use when the situation changes?
That question has become harder to avoid as AI tools grow more capable. Students can use them to draft, solve, summarize, and revise. Those uses may be appropriate in some assignments. But a finished product alone may not tell us what the student understands or can do.
Assessment is often treated as a way to measure learning at the end of instruction. It can also shape what students practice along the way. If a course rewards only correct answers or polished products, students may focus on producing those outcomes. If assessment asks them to make decisions, explain their reasoning, respond to new information, and revise their work, it can help make learning more visible.
In the near future, this may mean adding evidence to assignments we already use. A student might submit a report along with a short explanation of one important decision. A group might present its recommendation, then respond to questions about the evidence and trade-offs behind it. In a technical lab, students might document what they tested, what they learned from each result, and how they changed their approach when the first fix failed.
AI can be part of these assessments. Students might ask it for possible solutions, then compare its suggestions with course materials or test them in a lab. The assessment can focus on how students evaluated the response and what they did with it. The goal is not to catch AI use. It is to understand the student’s learning, including how they use tools.
Over the next decade, colleges may also develop more flexible ways for students to show what they can do. A student might build a portfolio across courses, combine classroom work with workplace evidence, or demonstrate a skill through a project, simulation, or live performance. AI-supported systems could help organize evidence over time, making it easier to see how a student’s capabilities have developed.
That future raises practical questions. Who decides which evidence counts? How do colleges ensure assessments are fair across different students and learning environments? How can a portfolio show growth without turning every moment of learning into something that must be recorded and evaluated?
A portfolio or performance is not automatically better than an exam or paper. Each assessment needs to fit its purpose. A timed test may be useful for checking foundational knowledge. A written assignment may help students develop an argument. A simulation may let them practice responding to changing conditions. The important question is whether the assessment gives students a meaningful opportunity to demonstrate the learning the course intends to develop.
Assessment also needs to remain manageable. Faculty cannot add a lengthy oral defense to every assignment, and students need clear expectations about what they are being asked to demonstrate. Sometimes one well-chosen question, a brief reflection, or a short explanation during a lab can provide useful evidence without expanding the workload substantially.
Instructors can start by asking: What would convince me that a student can use this learning beyond the assignment? The answer might be a student’s ability to explain a concept in their own words, apply it in a new scenario, troubleshoot a problem, or identify when they need more information. Those are different kinds of evidence, and a course may need several of them.
Assessment will continue to change as AI becomes part of education and work. The opportunity is to use that change to make our expectations clearer. Students should know what they are learning, what they are expected to do, and how they can show their growing capability. Faculty should have evidence that helps them respond to students’ needs and make informed judgments about learning.
If colleges assess knowledge and capability in a variety of ways, what should every graduate be expected to know and do? That is the question for the next post.
Continuing the Conversation
Series 1: AI Is Exposing Existing Problems ✓ Completed
Series 2: What We Do About It ✓ Completed
Series 3: Cultivating Human Thinking in an AI World ✓ Completed
Series 4: Learning Alongside AI ✓ Completed
Series 5: Preparing Students for an AI World ✓ Completed
Series 6: The Purpose of Education in an AI World
Series 7: The Future of Higher Education
Current Post (2 of 8): What Should Assessment Become?
Next Up: The New Curriculum