College admissions essay reviewers say that certain words, such as “tapestry”, appear more frequently in submissions now and can be red flags that AI was used
Context & Ripple Effects
Reviewers’ attention to recurring, polished vocabulary is an early qualitative signal that generative writing is changing how admissions essays are read. It aligns with a later analysis finding that a measurable share of student papers showed signs of AI assistance, making authorship assessment a practical issue rather than a purely hypothetical one.
The challenge is not simply identifying machine-written prose: subsequent reporting found that generic student writing can be falsely flagged by AI detectors. That makes vocabulary-based suspicion useful only as a prompt for closer review, not as proof.
First-order effects
- Admissions readers may give essays containing increasingly common AI-associated phrasing additional scrutiny, raising the importance of applicants being able to explain and stand behind their writing.
- Applicants using generative tools face a less predictable review process because stylistic cues—not just disclosed tool use—can shape perceived authenticity.
Second-order effects
- Admissions offices will be pressured to define how reviewers should treat AI-style signals so that inconsistent individual judgments do not become a de facto screening rule.
- Detection vendors and guidance providers face a credibility problem: tools and heuristics that flag conventional or generic prose can create avoidable disputes for students and schools.
Third-order effects
- If AI-assisted drafting becomes routine, admissions evaluation is likely to move away from treating the essay as a standalone proof of authorship and toward processes that test context, voice, and applicant ownership.
- This is part of an emerging authenticity arms race: as schools look for AI signals, applicants may alter prose to avoid them, making stylistic detection less durable and potentially less equitable.
The trend: Generative AI is turning high-stakes writing from a finished work product into an increasingly contested signal of individual authorship.