The UK Just Killed the College Essay. Is the US Next?

For decades, the personal statement has been the centerpiece of college admissions on both sides of the Atlantic. In the UK, it was a single 4,000-character free-form essay. In the US, it is a 650-word Common App personal essay plus school-specific supplements.

Then AI happened.

In July 2024, UCAS announced it would replace the traditional personal statement with three structured short-answer questions for 2026 entry. The old format --- a blank page and a word count --- is gone. In its place: targeted prompts that are harder for AI to game and fairer for students who lack essay-coaching support.

The question every American applicant should be asking: will US colleges follow?

---

What UCAS Actually Changed

Starting September 2025 for 2026 entry, UK applicants no longer write a single personal statement. Instead, they answer three structured questions:

1. Why do you want to study this course or subject? --- Demonstrate knowledge of your chosen field, what sparked your interest, and what you have done outside the classroom to develop that interest.

2. How have your qualifications and studies helped you prepare for this course? --- Showcase relevant and transferable skills from your academic work.

3. What else have you done to prepare outside of education, and why are these experiences useful? --- Cover extracurriculars, community work, summer programs, or hobbies that connect to your chosen course.

The total character limit remains 4,000 (including spaces), with a minimum of 350 characters per question. Students can allocate the remaining characters however they choose.

Why the Change?

UCAS cited two driving factors:

• Equity. Research found that students from disadvantaged backgrounds struggled more with the open-ended format, often lacking access to the essay coaching that wealthier applicants take for granted. Structured questions level the playing field by telling every applicant exactly what to address.

• AI. While UCAS has not stated this as the sole reason, the timing is telling. Open-ended prompts are exactly what large language models excel at. A 4,000-character free-form essay is trivially easy to generate with ChatGPT. Three specific questions with minimum character counts per response are harder to automate convincingly --- especially when each question demands different evidence.

---

Why the Open-Ended Essay Is Vulnerable to AI

The traditional college essay format has a fundamental design flaw in the age of AI: it rewards polished prose on broad topics, which is exactly what LLMs produce best.

Give ChatGPT a prompt like "Write a 650-word essay about a challenge you overcame" and you will get a competent, well-structured narrative. It will have a hook, a turning point, and a reflective conclusion. It will sound like a good essay. It just will not sound like you.

Shorter, more specific prompts are harder to fake because they demand:

• Particular facts about a named program, course, or experience that the model cannot invent
• Connections between disparate pieces of the applicant's life that require genuine self-knowledge
• Brevity that forces precision --- AI-generated text tends to be verbose and generic when constrained to 350-character minimums across multiple questions

The UCAS restructuring does not eliminate AI risk. But it raises the cost of using AI from "paste and submit" to "edit heavily and still risk sounding generic."

---

US Schools Are Already Experimenting

American universities have not abandoned the essay, but several are adapting in ways that mirror the UCAS logic.

Princeton: The Graded Paper Requirement

Princeton asks applicants to submit a graded written paper, preferably from an English or history course, completed during the last three years of high school. The paper must include the teacher's grade and any comments. Creative writing is not accepted.

This is a direct countermeasure to AI. You cannot generate a graded paper because it requires a teacher's evaluation attached to a specific assignment. It is a verified writing sample in a way that a personal essay can never be. Princeton's JSON data confirms the school is L1/D0/E1 --- they discourage AI use through statements like "an essay generated by an AI platform is unlikely to be as rich and nuanced as a student's own words" but rely primarily on the honor system and this graded paper mechanism rather than formal disclosure.

Amherst: Graded Paper as a Verification Tool

Amherst College also accepts graded papers as part of the application, joining a growing group of about 40 Common App member colleges that accept or encourage graded writing samples. Amherst itself has no explicit AI policy (L0/D0/E0), emphasizing "original, personal responses" without formal enforcement. The graded paper serves as an implicit authenticity check: admissions officers can compare the polish of an applicant's personal essay against the reality of their classroom writing.

Duke: Eliminating Numerical Essay Scores

Duke took a different approach entirely. In 2024, Dean of Undergraduate Admissions Christoph Guttentag announced that Duke would stop assigning numerical ratings to admissions essays. The reason was explicit: "We're just no longer assuming that the essay is an accurate reflection of the student's actual writing ability."

Previously, Duke scored applicants on a 1-5 scale across six categories, including essays. Now, essays are evaluated holistically for content and fit rather than writing quality. Duke's data (L2/D0/E1) confirms a policy that allows limited AI use at the institutional level, while their Law School program goes to L4/D0/E2 with an outright prohibition.

For the 2025-2026 cycle, Duke also introduced an optional AI-focused essay prompt, asking students to describe a situation in which they would or would not choose to use AI.

Virginia Tech: AI on the Other Side of the Desk

Here is the most striking case. Virginia Tech (L0/D0/E0) has no explicit AI policy for applicants --- but it uses AI to review essays. Starting in 2025-2026, an AI reviewer works alongside each human reader to "confirm the human reader essay scores, not make any admissions decisions."

This creates a notable policy gap: the university deploys AI to evaluate your essay but has not formally addressed whether you can use AI to write it. Virginia Tech's general academic integrity guidance suggests unauthorized AI use "may fall under several definitions of academic dishonesty in the Undergraduate Honor Code" --- but nothing in the admissions-specific policy draws that line.

---

The Common App Has Not Changed

Despite the AI upheaval, the Common App has made no structural changes to its essay format. The 2025-2026 cycle uses the same seven prompts as the previous year. The word limit remains 650. The format remains an open-ended personal essay.

The Common App does treat AI-generated content as fraud, and every applicant must e-sign an attestation that their work is original. But the essay format --- broad prompts, generous word counts, no verification mechanism --- remains exactly the kind of template that AI handles best.

Compare this to UCAS's three-question structure and the contrast is stark.

---

The L-Dimension Problem: Most Schools Have No Policy at All

Our data shows that ~67% of the 150+ schools we track are L0 --- meaning they have no explicit admissions AI policy. Schools like Harvard (L0/D0/E1), MIT (L0/D0/E0), Stanford (L0/D1/E1), Columbia (L0/D0/E0), and Amherst (L0/D0/E0) are relying on the essay format itself, combined with the Common App fraud policy, as their primary defense.

If the essay format changes --- as UCAS has demonstrated it can --- the entire policy landscape shifts. An L0 school that depends on the authenticity of the essay as a gatekeeping mechanism loses that mechanism the moment the format no longer serves that purpose.

The schools that have already adapted (Princeton with graded papers, Duke with holistic evaluation, Virginia Tech with AI-assisted scoring) are implicitly acknowledging what UCAS made explicit: the traditional essay is no longer a reliable signal of a student's authentic voice.

---

What Might Come Next

If American colleges follow the UCAS trajectory, here is what the landscape could look like:

• Shorter, more specific prompts. Rather than "Tell us about yourself in 650 words," expect questions that demand particular details about named programs, courses, or experiences. Harder for AI to generate, easier for admissions officers to verify.

• Timed writing samples. Some schools may introduce proctored or timed essay components, similar to how the SAT Essay once functioned. A timed response in a controlled environment eliminates AI assistance entirely.

• Portfolio-based submissions. Expanding the Princeton model: multiple verified writing samples from different contexts, with teacher grades and comments attached. This shifts the burden from one unverifiable essay to a body of verified work.

• Interview-based assessment. Real-time conversation remains the hardest thing for AI to replicate. Schools that expand interview programs gain an authenticity signal that no essay format can match.

• Hybrid approaches. A combination of structured short-answer questions, one verified writing sample, and an optional open-ended essay could give admissions officers multiple data points to triangulate authenticity.

---

The Counterargument: What Essays Do That Structured Questions Cannot

Not everyone agrees the essay should change. The strongest argument for keeping the open-ended format is that it reveals something structured questions cannot: voice.

A 650-word personal essay gives an applicant room to be funny, vulnerable, surprising, or strange. It allows for narrative arcs, unexpected connections, and the kind of personality that does not fit neatly into three targeted prompts. Some admissions officers argue that the messiness of the open-ended essay is a feature, not a bug --- it rewards the students who take creative risks, not just the ones who follow instructions well.

There is also a practical argument: structured questions can be gamed too. Applicants (or their AI tools) can research the "right" answers to specific questions more easily than they can fabricate a genuine personal narrative. The open-ended format's strength is that there is no single correct answer.

The tension is real. But the UCAS decision suggests that, at scale, the equity and AI-resistance benefits of structured questions outweigh the expressiveness of the open-ended essay.

---

What You Should Do Right Now

Regardless of whether American colleges restructure their essays, the direction is clear:

1. Expect more verification. Graded papers, writing samples, and interview-based assessments are expanding. Build a portfolio of strong classroom writing now.

2. Read your target schools' policies. Use the GradPilot AI Policies Directory to understand where each school stands. Check the methodology page to understand what L/D/E codes mean and how we verify them.

3. Write essays that only you could write. The best defense against both AI suspicion and format changes is specificity. Name real people, real moments, and real details. AI cannot invent your life.

4. Prepare for structured questions. Even if the Common App does not change, individual schools are already moving toward shorter, more targeted prompts. Practice answering specific questions concisely.

---

Related Reading

• Do Colleges Check for AI? What Top 10 Universities Actually Say
• Do Colleges Use AI Detectors? The Turnitin Truth
• Why Turnitin Failed College Admissions: 15% Miss Rate + Replacements
• How We Classified 150+ University AI Policies
• How Much Are Colleges Spending on AI Detection Tools?