In this discussion, several users criticize Virginia Tech’s plan to use AI to grade admissions essays, arguing that if essays can be reliably graded by AI, they may not be worthwhile in the first place. Commenters question how VT can avoid the factual errors, inconsistent logic, and hallucinations common in AI systems, sharing personal experiences where different AI tools produced contradictory technical information. Others note that AI output is only as good as the underlying language model and that early expectations for AI were oversold. Additional replies joke that since many essays are written by AI anyway, it might as well grade them, while others pivot to broader frustrations about administrative bloat at the university and the reliance on volunteers to read essays. One commenter mentions a relative who used to volunteer as a VT essay grader and often criticized how the process was managed. ** Edited by chumley at 12/2/2025, 1:49:04 PM