AI Access And AI Advantage Are Not The Same Thing
Why this matters to Wen: higher-ed AI stories get shallow fast unless equity stays in the frame.
Applications analyzed in a 2026 study of LLM use in U.S. college admissions essays.
The study tracks how LLM-assisted writing spread from 2020 to 2024 and shows that access does not map cleanly onto equitable outcomes.
The Story
This admissions-essay paper is one of the clearest recent reminders that AI narratives become dishonest when they stay at the level of access alone. The researchers analyzed 81,663 applications and found that estimated LLM use rose sharply in 2024 across groups, with disproportionately larger increases among lower-SES applicants.
That might sound like a straightforward leveling story, but the harder finding matters more. Increased estimated LLM use was more strongly associated with declines in predicted admission probability for lower-SES applicants than for higher-SES applicants, even after controls. In other words, the tool may widen support in one direction while still failing to equalize outcomes in another.
For Wen, this matters because higher-ed marketing and communications increasingly talk about AI as a readiness signal. That story needs more honesty. Institutions do not just need shiny innovation language. They need messaging, policy, and student support that confront where AI might reproduce disadvantage even while expanding access.
Studio Wensday Angle
When you see a university AI story this week, ask two questions before admiring it: who benefits first, and who still pays the hidden cost? That frame will produce sharper strategy and better content than generic future-readiness claims.