What generic AI gets wrong with this prompt
This option looks warmer than Yale’s other two, which is exactly why generic AI does more damage here. Warmth invites vagueness, and vagueness is what the model produces best.
It polishes a community label anyone could claim
Write “my culture” or “my hometown” and generic AI will mirror the label back with better grammar instead of pressing on what the community actually is. The model treats your first framing as settled. But an opening paragraph that could belong to anyone who shares your background tells the committee the essay was easy to write.
Ask your AI — “Mark every sentence where I am the one doing something. How much of this essay is about the community, or about other people, rather than about me?”
It refuses to pick one thing gained and one thing contributed
Yale asks for the most valuable thing you gained and the most important thing you contributed — singular, twice. Generic AI hedges with three of each, because choosing requires judgment about your life it doesn’t have, and lists feel safer than commitments. Two firm choices in 500 words beat six soft ones every time.
Ask your AI — “State in one sentence each the single thing I gained and the single thing I contributed in this draft. If you can’t, where does the draft avoid choosing?”
Your contribution comes out as a member’s resume
Asked what you contributed, generic AI formats whatever roles and titles you gave it — treasurer for two years, organized the annual event — because that’s the shape your inputs arrive in. Tenure and titles describe membership. The question is about chosen acts, the ones that cost you a weekend or a hard conversation, and those don’t appear unless you put them in.
Ask your AI — “Which sentences describe a position I held, and which describe a specific act I chose to do? What did any of it cost me?”