Grad School AI Policies: MPH, MSW, and Counseling Rules

Grad School AI Policies: MPH, MSW, and Counseling Rules

Program AI policies across the health master's disagree with each other: there is no field-wide answer to whether you may use an AI tool on your personal statement, because the programs have written sentences pointing in four different directions — and most programs have written nothing at all. Every rule below was retrieved on August 11, 2026, and is quoted in the program's own words:

• Two ban assistance outright. Tennessee's MSSW says applicants "must not seek the assistance of Generative AI Tools like ChatGPT" and calls such use "academic dishonesty." Northern Iowa's MSW says "The use of generative AI is prohibited in crafting your personal statement."
• One asks for materials "unenhanced by the use of generative AI" — UAB's School of Public Health, on the page governing its standalone MPH.
• One permits AI in a narrow lane. Rutgers MSW allows it "only permitted to check grammar and sentence structure," and nothing else.
• One requires disclosure rather than abstinence. MSU Denver's MSW asks applicants to describe, in one to five sentences, whether and how they used generative AI, including the prompts.
• Two make you attest to authorship without naming AI at all — DePaul's counseling master's and Harvard's Chan School.
• The centralized platforms regulate responsibility, not review. SOPHAS, SocialWorkCAS, and GradCAS put the application on you and point at a code of conduct; none publishes an AI permission or an AI ban on its public applicant pages.

Seven programs and three platforms are a sample, not a directory. The point is not the count: a ban and a permission coexist inside the same profession and the same cycle, so there is no MPH answer, MSW answer, or counseling answer to inherit — only your program's answer to find. Below is every sentence we could verify, plus an honest reading of what each posture means for having anyone, human or machine, read a draft you wrote yourself.

Health master's AI policies, in one table

Every link is in the sources at the end.

Bans that reach assistance, not just generation

Three sentences in this set are wider than "do not submit AI-written text."

The University of Tennessee's College of Social Work publishes the widest one on its MSSW personal statement page, which asks for five to six double-spaced pages under headings and carries no application-cycle label:

Applicants must not seek the assistance of Generative AI Tools like ChatGPT. Use of a Generative AI Tool to complete this admission application constitutes academic dishonesty.

Read both sentences. The first regulates seeking assistance, which is broader than producing text — a tool that only comments on writing you already finished is still assistance you sought. The second supplies the consequence category up front, which very few admissions pages do. We are not going to narrow someone else's sentence on your behalf, so the practical consequence goes here plainly: if you are applying to the Tennessee MSSW, do not run that statement through GradPilot or any comparable outside AI review.

The University of Northern Iowa's MSW admissions page, which caps the statement at three pages, publishes the "crafting" variant:

The use of generative AI is prohibited in crafting your personal statement.

"Crafting" is the same verb problem readers meet in nursing school AI policies, where one DNP program uses it too. It is wider than generating and narrower than Tennessee's "assistance," and nothing on the page resolves where revision sits. The department separately maintains a standard against generative AI in its social work courses unless an instructor directs otherwise, so the posture is programmatic rather than incidental. Treat it as no-use unless the program tells you otherwise in writing.

UAB's School of Public Health carries the MPH version on the graduate application page:

Application materials such as statement of purpose & letters of recommendation should be in the author's own words, unenhanced by the use of generative AI, and citations used where necessary.

"Unenhanced" is the operative word, and it reaches revision. A review that returns critique is offered precisely so the next draft is better than the last one — which is enhancement, whoever types the improved sentence. Same conclusion: do not use an AI review on a UAB MPH statement.

A permission that is narrower than it looks

Rutgers is the only program in this sample that affirmatively permits AI, and the permission is a ceiling rather than a door:

The use of Artificial Intelligence (AI) tools is only permitted to check grammar and sentence structure. AI tools should not be used to write content about you, your motivations, or your experiences.

Two words matter: only and check. Grammar and sentence structure are mechanical properties. A rubric review is not that — it reads argument, evidence, and fit, which is the whole reason it is worth running. So the honest reading is the conservative one: a substantive AI review of a Rutgers MSW statement sits outside what that sentence permits, even though the same sentence would clearly allow a spell checker.

It is worth knowing why Rutgers writes so precisely. The same page sets the statement at "750-1,000 words in total" and tells applicants: "We do not offer individual interviews and often the personal statement will determine your success in the application process." A program that has removed the interview has made the essay the interview, and it is regulating the essay accordingly. The four questions behind that word count are worked through in our MSW personal statement guide.

Disclosure instead of prohibition

MSU Denver's MSW takes the fourth path. Its Summer/Fall 2026 personal-statement document — a set of short answers with individual word limits, saved in one file — ends with a prompt that is not about social work at all:

Briefly describe (1-5 sentences) whether you used Generative AI in the process of writing your personal statement and if so, how you used this tool and the prompts you used.

Notice what it asks for beyond a yes or no: how you used it, and the prompts you used. That is a records requirement disguised as an essay question, and it means deciding now what you will be describing later — reconstructing a prompt history after the fact turns a defensible process into a guess.

A disclosure regime is neither permission nor a trap. It changes the question from "am I allowed" to "can I describe what I did in five sentences and be comfortable with how it reads to the committee." Some processes survive that sentence easily. Others do not survive it at all.

One caution: the document above is the 2026 cycle. The 2027 walkthrough says the prompts and the case study publish when the application opens on September 2, so that wording — including the AI prompt — must be re-read when it appears. Prior-cycle wording is evidence of posture, never proof of the current rule.

Attestations that never mention AI

Two of the strongest sentences in this set do not use the word "AI" at all.

DePaul's counseling master's asks for a statement of no more than 1,500 words answering five questions, and requires the statement itself to carry an attestation:

I attest that this essay is an accurate reflection of my writing ability. This personal statement is my original work and has been written by me, [insert your name].

"An accurate reflection of my writing ability" is doing unusual work. Most attestations regulate authorship; this one regulates representativeness. A statement can be entirely your own words and still misrepresent your writing ability if the visible prose is not the prose you can produce — which is what a committee is implicitly asking when it weighs the essay.

Harvard's Chan School does the MPH version, alongside a credit rule most applicants read past: "All materials submitted as a part of your application are expected to be your own work," and "Please credit any sources you've used in your application, whether that source is a publication or a conversation." The same page caps the Statement of Purpose and Objectives at 600 words "although SOPHAS allows for a higher word count" — direct evidence that a platform ceiling never overrides a program's lower limit. Those mechanics are in our MPH personal statement guide.

Authorship attestations reach any process that would make the submitted words not yours, whatever produced them. That is why they belong in an AI reference even when they predate the technology, and it is the same logic behind the attestations undergraduate applicants sign.

What the platforms actually regulate

Applicants keep hunting for the SOPHAS rule or the SocialWorkCAS rule. The centralized services regulate responsibility, not review.

SOPHAS. Its release statement tells applicants they are "expected and required to take full responsibility for their application throughout the admissions process and required to abide by the SOPHAS Applicant Professional Code of Conduct and Admissions Code of Conduct," with a separate checkbox acknowledging the accuracy of the information provided. The substantive integrity terms live one link away, on ASPPH's SOPHAS code-of-conduct page, which places responsibility on the applicant and draws the boundary at falsification and plagiarism. That page did not serve to our fetcher on August 11, 2026, so treat our description of it as a browser-read summary and read the code yourself before you sign — we found no public AI permission and no public review ban in either document. SOPHAS lists the 2026-27 cycle as opening August 12, 2026; the live acknowledgement inside the application is the version that governs you.

GradCAS, which carries many counseling programs alongside PsychologyCAS and direct portals, uses parallel language: applicants are "expected and required to take full responsibility for their application throughout the admissions process and required to abide by the GradCAS Applicant Code of Conduct." No AI sentence appears on the public release page.

SocialWorkCAS is the honest gap here. Its applicant portal opens in early September, its Program Materials section is customized by each program, and the contractual release text is not exposed publicly beforehand. We could not read it, so we are not going to characterize it. Read the certification inside your own application and treat it as governing.

The structural consequence matters more than any single clause: because Program Materials belong to each program, the essay rule belongs to each program too. There is no universal SOPHAS or SocialWorkCAS prompt, no universal length, and no universal AI rule.

What "review of your own writing" means under each rule

This is the question the whole page exists to answer.

Two rules apply across every row. A university-wide or undergraduate policy is not automatically your graduate program's policy — the mis-attribution trap documented across programs whose AI rules contradict their own university's. And a silent page is not a permissive one; across higher education the majority position is no published rule at all, which resolves nothing in your favor.

Finding your program's sentence in five minutes

1. Start on the page carrying your essay instructions, not the university's academic-integrity page. Every sentence quoted above sits beside the writing instructions it governs.
2. Name what the sentence regulates — generating text, crafting the draft, enhancing materials, seeking assistance, or completing and submitting the application. Five different scopes; the last two are the broadest.
3. Check the department, not the university. Public health, social work, and counseling programs at one institution publish separately and contradict each other regularly.
4. Read the certification inside the portal before you submit and treat it as governing, especially where the public page is silent.
5. Re-check late. MSU Denver's 2027 prompts publish at the September 2 open, and Arkansas's MSW page says its 2027 instructions changed outright.
6. When the rule is unclear, ask the program in writing. An admissions office's email answer beats any inference, including ours.

Where a GradPilot review fits, and where it does not

GradPilot reads writing you already wrote. It does not draft, does not rewrite, and does not hand back replacement prose you could paste in. What comes back is rubric-scored critique tied to your own sentences, plus authenticity signals that describe how the text reads.

Three limits belong in the same breath. An AI-sounding flag is a signal about how a passage reads, not proof of how it was made — the false-positive problem is mathematically unavoidable and lands hardest on writers whose English is a second language. Scores are revision signals, not admissions predictions, and they vary a little between runs on the same text. And no review certifies compliance with any program's rule; that determination is yours, from your program's current words.

Check the policy first, then the rubric. Where the program permits review, the MPH personal statement rubric, the MSW personal statement rubric, and the counseling statement of purpose rubric show the criteria before you pay, and the graduate statement of purpose review runs your draft against the one you choose with your program's exact prompt pasted above it. Two Quick Reviews are free each day; your first Full Review is $5, once per account, and Pro is $50 for ten Full Reviews. If you have only ever pasted a draft into a chatbot and been told it was excellent, the reason that feedback was useless has a name: gradeglazing.

Applicants to the Tennessee MSSW, Northern Iowa MSW, and UAB MPH should not use it for those statements, and the same caution applies anywhere a program restricts outside or AI-assisted feedback.

The short version

Seven programs, checked August 11, 2026, produced two assistance-level bans, one "unenhanced" rule, one grammar-only permission, one disclosure requirement, and two authorship attestations that never say "AI" — while all three centralized platforms regulate responsibility and code adherence instead. Find your program's own words, name exactly what they regulate, read the certification you are signing, and check again before you submit. Where the rule is a ban, the answer is no, and no rubric is worth an integrity finding.

More from this wave

• The MSW personal statement under SocialWorkCAS
• The MPH personal statement for the SOPHAS cycle
• Counseling statements: lived experience without the therapy-session register

Related reading

• Nursing school AI policies: 7 schools, rules both ways — the same treatment for NP, DNP, and accelerated BSN admissions.
• Law school application AI policies — each school writing its own sentence on its own timeline.
• When the same school publishes different AI rules — why a university policy is not your program's policy.
• Most colleges publish no AI policy at all — what silence does and does not mean.
• What you are actually signing in an AI attestation — the pledge language, decoded.
• The statement of purpose question bank — the underlying questions graduate prompts keep asking.

Sources

All retrieved August 11, 2026.

• University of Tennessee MSSW personal statement — assistance ban and academic-dishonesty sentence; five-to-six-page format; no cycle label on the page
• University of Northern Iowa MSW admissions — "crafting" prohibition; three-page maximum
• UAB School of Public Health graduate application — "unenhanced by the use of generative AI"; Statement of Purpose and Objectives requirement
• Rutgers MSW application process and deadlines — grammar-and-sentence-structure permission; 750-1,000 words; no individual interviews
• MSU Denver MSW personal statement prompts, Summer/Fall 2026 — generative-AI disclosure prompt; 250-word short answers
• MSU Denver 2027 MSW application walkthrough — 2027 prompts and case study publish at the September 2 open
• DePaul counseling MA/MEd admission requirements — attestation sentence; 1,500-word five-question statement
• Harvard T.H. Chan application requirements — own-work and source-credit rules; 600-word MPH limit below the platform ceiling
• SOPHAS release statement — responsibility and code-adherence acknowledgement
• ASPPH SOPHAS Code of Conduct — applicant responsibility and the falsification/plagiarism boundary (browser-only; did not serve to our fetcher on this date)
• SOPHAS application cycle dates — 2026-27 cycle opening date
• SOPHAS Program Materials — requirements and additional essays vary by program
• SocialWorkCAS Program Materials — program-customized requirements
• SocialWorkCAS release statement help page — contractual text not exposed publicly before the cycle opens
• GradCAS release statement — responsibility and code-adherence language; no AI clause
• University of Arkansas MSW admissions — 2027 instructions changed; case study publishes October 1, 2026

Rules move, and pages move with them. Re-check your program's current wording and the certification inside your own application before you submit. The program's live text governs, not this page.