Will Students Explain Their Thinking More Often in the AI Age?

A polished essay or completed project can offer valuable evidence of learning. As generative AI becomes more capable, educators also need to understand the decisions, knowledge, and reasoning that produced the work. That is renewing interest in conversations, demonstrations, and portfolios.

These approaches have a long history. What is changing is the urgency of deciding which evidence can support a trustworthy judgment about learning, and how new technologies might help collect it without making assessment less fair or more burdensome.

Assessment is becoming a design question

A July 2026 Stanford and ETS discussion of responsible assessment argues that final products alone may no longer represent the capabilities an assessment intends to measure. It points toward evidence gathered across tasks and time, with attention to validity, fairness, transparency, and human oversight.

A 2026 study of electronic portfolios and interactive oral assessments examined the perceptions of 85 undergraduate biopsychology students. Participants viewed the formats differently, and students with higher support needs rated portfolios more positively. This was a study of experiences and perceptions in one university subject, not proof that either format produces better learning across school settings.

Ask students to make a decision visible

A teacher could invite a student to explain why they chose one method, what changed between two drafts, or which evidence might challenge their conclusion. A brief conversation can complement the work already submitted and help the teacher identify the next instructional step. Difficulty answering one unexpected question should not be treated as proof that a student used AI improperly.

For example, after a design project, students might explain one choice and one revision using their sketches. The teacher can then ask how the design would respond to a changed constraint. That provides a different kind of evidence from asking the student to repeat a prepared presentation.

Build fairness into the format

Speaking confidently is not the same as understanding a subject. Oral assessment can introduce barriers related to anxiety, disability, language, processing time, or unfamiliar expectations. Students need clear criteria, opportunities to practise, and appropriate accommodations.

Where the goal is scientific reasoning, a diagram, written explanation, or supported conversation may provide relevant evidence. Where oral communication is itself the goal, it should be taught and assessed explicitly. The format should serve the intended capability.

Let a few work samples tell a learning story

A portfolio need not become an archive of everything a student has produced. A small selection can show an initial attempt, feedback, revision, and a reflection explaining what changed. Students can learn to choose evidence and describe its significance.

Schools should also consider whose time the process consumes. A short, purposeful discussion or a few carefully selected work samples may be more manageable than adding lengthy recordings and reflections to every assignment. Recording introduces additional privacy and retention questions that require clear decisions.

Keep assessment a source of understanding

The strongest reason to hear students explain their thinking is the information it gives teachers and learners. It can reveal a misconception, an unexpected connection, or a productive next question.

As assessment evolves, schools can ask a practical question: what combination of evidence would help us understand this student's learning more accurately, fairly, and usefully?

Illustrative image generated with AI.