In a 2025 study in Medical Education, 28 educators evaluated medical students’ written reflections alongside versions whose reflective sections were generated by AI. They could not reliably distinguish the two. The small, single-school study raises a larger question for colleges: What does polished writing reveal about the person behind it?

As a former admissions professional at Wharton and a college admissions advisor, I approach this as a question about evidence. Colleges need to understand how applicants make decisions, respond to uncertainty and learn. An elegant essay can contribute to that picture, but it cannot carry the assessment alone. And identifying these capacities at admission creates a responsibility to develop them afterward.

In my earlier Forbes article on AI fluency, I argued that students need to evaluate AI output and explain when they override it. That requires colleges to make judgment visible throughout the educational experience.

When I interviewed Kelly Dore, co-founder and vice president of science and innovation at Acuity Insights, she described a problem that predates generative AI: personal statements shaped by extensive outside assistance.

“So what it became is really a village statement of those you had around you rather than a statement that told them anything about you,” Dore said.

Dore co-created Casper at McMaster University with Harold Reiter. The situational judgment test asks applicants to explain how they would respond to difficult scenarios. Acuity’s standard format includes 11 scenarios with typed or recorded video responses. According to the company, trained human raters score the responses, with a different rater evaluating each scenario.

A public sample involves a classmate who stops contributing to a group project because of a medical condition. A follow-up changes the context: What if the classmate had already contributed substantially? The reasoning becomes observable through the applicant’s explanation of whether that information changes the decision.