Asked to Write Your Own Med School Rec Letter? A Guide

Asked to Write Your Own Medical School Letter of Recommendation? A How-To (and the Ethics)

"I'd be happy to sign a letter for you — just send me a draft."

If a physician, professor, or supervisor has said some version of this to you, you are not in an unusual situation, you have not been asked to do something unspeakable, and you also have not been handed a freebie. You have been handed the hardest writing assignment in the application: a document that must sound like someone else, evaluate you credibly, and coexist with a personal statement written in your actual voice — inside an application system that increasingly says the letter must be the recommender's own work.

This guide covers both halves honestly: the ethics (what is actually allowed, by whom, in writing) and the craft (how to draft a letter that a busy attending will genuinely adopt — and that will not read like you wrote it). It is the applicant-side companion to our guide for the people writing medical school letters, and everything here applies double at the residency stage, where attendings request drafts even more routinely — see the ERAS letters of recommendation guide for that system's rules.

How Common Is This, Really?

Common enough to have two decades of forum history. Student Doctor Network threads on the subject run from "doctor asked me to write my own letter of recommendation" in the mid-2000s to "Writing my own LOR" in 2025, with the community consensus hovering around "awkward, common, handle carefully." Outside medicine, a ProFellow poll found 79% of respondents had at some point been asked to draft their own recommendation letter. The request usually comes from the writers who would otherwise be your best letters — busy clinicians and PIs who know your work well and your deadline pressure better.

So no, you are not being singled out. But frequency is not permission, and the permission question has real answers. (This post is the medical-specific playbook; if your request came from a professor for graduate school, or an employer, the cross-field version is our general guide to being asked to write your own letter of recommendation.)

The Ethics: What the Rules Actually Say

There is no single AMCAS-wide rule that says "applicants may not draft their own letters." What exists instead is a patchwork — and the patchwork is tightening.

• The FERPA waiver you sign is a statement about authorship. When you waive your right to view your letters (as nearly all applicants do, and should), you are telling schools the letter is the writer's independent, confidential assessment. A letter you wrote yourself and never "saw" makes that waiver a technicality at best.
• Some application services are explicit. CASPA's applicant agreement bars applicants from influencing evaluator content, and CASPA has said it may run AI-detection tools on submitted materials — we covered the PA-side enforcement posture in our CASPA AI detection analysis. If you are a PA applicant, treat "draft it yourself" as materially riskier than an MD applicant should.
• Some universities ban it outright. Stanford's graduate admissions terms state that "drafting, writing, translating, or submitting your own reference, even if asked to do so by a recommender" is a violation — grounds for rejection or rescission. That is a graduate-school policy, not an AMCAS one, but it tells you where institutional sentiment is headed, and several schools apply similar language to all application materials. Our international students LOR ethics guide documents how actively U.S. universities verify letters when they suspect applicant authorship.
• The letter must be submitted by the writer, from the writer. Every major service — AMCAS, AACOMAS, TMDSAS, CASPA — routes letters through writer-side portals. TMDSAS additionally requires letterhead, the writer's contact information, and a current-cycle date. You drafting content is one thing; you submitting it is fraud everywhere.

The honest ethical ladder

From clearly fine to clearly not:

1. Clearly fine: giving your writer a brag sheet, CV, deadline list, and "here are three moments you might remember" bullets. This is not just allowed — the AAMC's own letter-writer guidance assumes writers work from applicant-provided context, and every prehealth office recommends it.
2. Defensible, with conditions: writing a first draft at the writer's explicit request, which the writer then genuinely reviews, edits, and owns. The conditions matter: the final assessments, comparisons, and signature must be theirs, and they must submit it.
3. Risky: a verbatim ghostwrite the writer signs unread. If the school ever compares the letter to your essays — and some schools run AI and authorship screening on application materials — you wrote the evidence against yourself.
4. Prohibited everywhere: submitting the letter yourself, inventing observations the writer never made, or fabricating the relationship. This is application fraud, and rescission clauses cover it even after matriculation.

The cleanest move is to try to convert the request down the ladder. Reply with: "I'd feel more comfortable sending you a detailed bullet-point summary of what we worked on together, plus my CV and personal statement — would that work?" A meaningful fraction of writers were asking for exactly that and just phrased it lazily. If the writer insists on a full draft, rung 2 is where you operate — and the rest of this guide is how to do rung 2 well.

The Craft: How to Draft a Letter You Didn't Have to Write

The central failure mode of self-drafted letters is not ethics — it is voice. Committees read your personal statement, your activity descriptions, and your letters in one sitting. A self-drafted letter tends to fail in three detectable ways:

1. It echoes your personal statement. Same anecdotes, same phrasing, sometimes the same sentence rhythms. If your essay says you "learned that medicine is practiced at the intersection of science and story," and your letter praises how you "work at the intersection of science and story," a reader needs no software to conclude one author wrote both.
2. It knows things only you know. A letter that narrates your internal motivations ("she realized in that moment that she wanted to serve underserved communities") is written from inside your head — a vantage point your supervisor does not have. Writers describe observable behavior; only applicants describe feelings.
3. It is too kind in the wrong ways and too timid in the right ones. Self-drafters pile on warm adjectives (which committees discount) and shy away from the concrete superlative comparisons (which committees value most — being called "the best" was one of only three letter features that predicted medical school performance in the Kirch 2014 study) because writing "she is the strongest student I have supervised in a decade" about yourself feels unbearable. The result reads warm, vague, and weak.

Work against all three deliberately:

• Interview your recommender first, even for ten minutes. Ask: "What do you actually remember me doing well? Who would you compare me to?" Draft only from what they said and saw. This single step moves the letter's content from your head to theirs — which is both the ethical point and the craft fix.
• Write the relationship statement from their vantage. Duration, capacity, direct observation: "Maria worked under my supervision as a scribe for 16 months, roughly 300 clinical hours." Structure the body as situation, behavior, consequence — the pattern the AAMC guidelines prescribe.
• Use different anecdotes than your personal statement. Nothing in the letter should retell an essay story. If your best material is already in the essay, that is the interview's job: surface moments you forgot.
• Put in the comparison you would never say out loud — flagged for their edit. Write it in brackets: "[If accurate: among the N students I have supervised, she is one of the strongest because…]". You are not asserting it; you are making it easy for the writer to assert it honestly. This also forces the one conversation that makes the whole exercise legitimate.
• Keep the tier language calibrated and the hedges out. One page, 400-600 words, no "mostly," no "improving," no restated grades — the same rules that bind any medical school letter writer.
• Do not use ChatGPT to generate it. A self-drafted letter is already carrying authorship risk; an AI-drafted self-drafted letter compounds it with detectability, and generic-praise output is exactly what committees discount. AI-written letters get noticed for craft reasons before detection tools ever run.
• Hand over an editable document and explicitly invite changes. "Please edit anything that doesn't match your view — especially the comparison in brackets" is the sentence that keeps you on rung 2. A writer who changes five sentences has made the letter theirs.

Before It Ships: The Two-Document Read

Whatever you draft, someone other than you should read the letter and your personal statement side by side, looking for echoed phrases, repeated anecdotes, and a suspiciously identical voice — because that is exactly the reading an admissions committee will perform. Your prehealth advisor can do it. So can we: GradPilot reviews draft recommendation letters against what committees actually reward — evidence density, comparison quality, hedge detection — and runs an AI-detection check on the draft, in minutes. We review the letter you are drafting; we never write it. If the letter comes back reading like your essay's twin, fix it before a committee is the one to notice.

The uncomfortable request your recommender made is survivable — applicants survive it every cycle. The ones who get burned are the ones who treat it as a shortcut instead of what it actually is: writing the most disciplined document in their application, twice removed from their own voice, and then giving it away.