A new analysis of tens of thousands of admissions essays at a selective college found rising use of homogeneous language after AI tools became available in 2022, with an observed pattern tied to socioeconomic status. The reporting centers on low-income applicants being more likely to submit AI-generated essays. The findings land as admissions offices expand AI guidance and as institutions face scrutiny over how applicants use generative tools to manage time and language demands. The study’s core claim is that AI use is not evenly distributed and may be altering the linguistic signal readers rely on. For higher ed professionals, the immediate implication is operational and policy: schools may need clearer AI-use definitions, stronger review protocols for essay inputs, and equity-focused assessments of how AI affects reading comparisons across applicants.