When AI Can Write the Essay, What Are Universities Really Evaluating?
by BrainStream Chief Marketing Officer Ya-Hui (Laura) Kao
University application season is here. Students are working on personal statements, supplemental essays, and other application materials designed to show who they are beyond their grades.
But this year, there is a question that is becoming harder for universities to answer:
How much does a finished essay actually tell us about the student who submitted it?
Generative AI can now help students brainstorm, organize ideas, improve their writing, and rewrite entire sections. The final essay may look polished and personal, but it can be difficult to know how much of the thinking and writing came directly from the student.
This is not simply a question of whether a student used AI.
A student who uses AI to improve grammar is very different from one who asks AI to write an entire essay. Yet when admissions officers only see the final product, that distinction can be difficult to see.
And the challenge goes beyond university applications. The same issue is appearing in classrooms, where essays, assignments, and take-home projects have traditionally been used to measure what students know and what they can do.
That may mean education needs to look beyond the final answer.
Instead of asking only, “Did the student produce this?”, we may need to ask:
Can the student explain it?
Can they defend their ideas?
Can they apply what they learned in a new situation?
Can we see how their thinking developed?
AI is not going away, and education does not need to treat every use of AI as a problem. Students will enter a world where AI is part of how people work and communicate.
The bigger challenge is making sure that technology supports learning without replacing the evidence of learning.
As universities rethink what application essays and student work can tell them, one thing may become increasingly important: the process behind the work, not just the final product.
Supporting Reading
Stanford University, A Roadmap for Responsible Student Assessments in the AI Era
