How Margin Uses AI

The short answer

Margin uses AI to deliver feedback — but it never uses AI to write your essay. It asks questions, points out what's working and what isn't, and helps you sharpen your own thinking. The words, the ideas, and the decisions are always yours.

It doesn't write your essay.

This is the one rule every part of Margin is built around: it never generates essay content for you to submit. When Margin sees a weak sentence, it doesn't rewrite it — it tells you what's not working and why, and leaves the fixing to you. When it sees a missed opportunity, it points to it — it doesn't fill in the gap itself.

The words stay yours. Your thinking gets sharper.

It's built to protect your voice, not flatten it.

Generic AI writing has a recognizable problem: it tends to smooth every voice into the same polished, slightly hollow register. Margin is built to catch exactly that, not produce it. It actively looks for places where an essay drifts away from a student's own established voice — a paragraph that suddenly sounds more formal, more "thesaurus," more like it's trying to impress than communicate — and flags it directly, the same way a real reader would notice when someone stops sounding like themselves.

This is about more than detecting AI. It's about making sure an essay sounds unmistakably like the one person who actually lived it.

It teaches — it doesn't just edit.

Margin doesn't hand back a score and move on. It works the way a real mentor does: asking what you actually meant by a passage, naming the single most important thing to work on next instead of overwhelming you with everything at once, noticing real growth across drafts, and always leaving you with something to think about, not just something to fix.

A real teacher designed the judgment behind it.

Margin's feedback is generated by AI, but the standards it applies — what makes an opening work, what admissions readers actually notice, when a reflection feels earned versus stated — were designed around real teaching principles developed over nearly two decades of working with student writers, not invented by a general-purpose chatbot. Every response is shaped by that same judgment, every time.

On AI detectors

We won't tell you Margin is "AI-proof" — no honest company can promise that, and we don't trust ones that do. Here's the more useful truth: AI detectors are inconsistent, and research has shown they disproportionately misflag writing from non-native English speakers, sometimes at dramatically higher rates than other students' writing. That's not a reason to fear them; it's a reason not to build your defense around beating them.

The actual, durable protection is a genuinely authentic essay — one so specific to your own experience, your own voice, and your own thinking that it couldn't have been written by anyone else. That's what Margin is built to help you produce, and it's a stronger safeguard than any detector score.

Human review, for real safety concerns

Margin is not a crisis service, a mental health provider, or a substitute for a trusted adult. If something a student writes suggests a genuine safety concern, Margin can offer supportive resources — but any decision with real-world consequences is always made by a person, never by AI alone. See our Student Safety Statement → for the full explanation.

In short

Margin isn't selling AI. It's offering real writing feedback, whenever you need it — a writing mentor's judgment, available any hour, from anywhere, for as many drafts as it takes.