Academic Integrity

Drafting and refining your own work. The legitimate use cases we care about:

• You've drafted something yourself and want it tighter, clearer, or in a different voice.
• You're a non-native English writer and want fluency help without losing your meaning.
• You used AI assistance to brainstorm or outline and want the final prose to read like you, not like a model.
• You need real, verifiable academic sources for claims you're making (Find Citations is backed by OpenAlex with DOIs verified against Crossref — no hallucinated papers).
• You want a second opinion on whether your draft reads as too uniform or templated.
• You're translating something you wrote in another language and want a faithful version that keeps your register.

In every one of these, the substantive thinking is yours. Penshift sharpens what's already there.

We are not an essay generator. We are not a tool for misrepresenting authorship. The cases below are not what we built Penshift to do, and using it for them is a breach of these guidelines and of our Terms of Service.

• Generating a complete piece of work from a prompt and submitting it to a course, journal, employer, or grant body that requires original human work, without disclosure.
• Running an AI-drafted piece through Penshift's humaniser to launder it past a detector your institution is using to enforce its integrity policy. Even where Penshift's output passes a detector, the underlying work is still AI-generated, and submitting it as your own without disclosure is misconduct.
• Writing on behalf of another student or researcher who is meant to be the author.
• Producing fabricated quotations, fake citations, or invented data and dressing it up as real research.

Penshift Chat is configured to refuse the first of these directly. The other tools don't know your context, so the responsibility for using them ethically sits with you.

Our humaniser is good at producing prose that reads like a human wrote it. On simpler classifiers it usually drops well into the human range. On stricter commercial detectors (Originality.ai, Copyleaks, GPTZero in academic mode, Turnitin) the results are mixed: some outputs pass, others still flag. We don't promise to bypass any specific detector and we don't think anyone honest can.

If detector evasion is your only goal, you're using the wrong tool. Penshift is built to help AI-assisted writing read like human writing. Better detector scores happen as a side effect, not as the headline claim.

More importantly: a detector pass is not an integrity pass. Most universities and journals now require AI-use disclosure regardless of whether a detector flags the work. Whether a detector caught you isn't the question. Whether you complied with your institution's policy is.

Different institutions have different rules about AI assistance:

• Some allow AI for brainstorming and editing but not drafting.
• Some allow it for any purpose if disclosed.
• Some prohibit AI assistance entirely.
• Many require you to declare on submission whether and how AI tools were used.

Penshift has no visibility into your specific course's, journal's, or employer's rules. Before you use Penshift on anything you intend to submit, check the policy. If disclosure is required, disclose. If AI use is prohibited, don't use Penshift on that piece.

If you're unsure, ask your instructor, supervisor, editor, or institution's academic integrity office. “Penshift didn't tell me to disclose” is not a defence.

Coursework essays and exam responses.If your course prohibits AI use, don't use Penshift on the submission. If your course requires disclosure, disclose. If you used Penshift to refine prose you yourself wrote, that's usually within bounds, but the rule is your institution's, not ours.

Personal statements and applications.Most admissions offices allow editing assistance and require that the applicant's voice and content be authentic. Penshift can help refine your draft. It cannot honestly write the statement for you. If a friend, tutor, or AI substantively wrote your personal statement, it isn't your personal statement.

Research papers and theses. Most journals now require AI-use disclosure in cover letters or methods sections. Many specifically ask whether AI generated text or analysed data. Penshift assistance with prose typically falls under these disclosure rules. Penshift cannot be listed as an author of a paper.

Cover letters and job applications.Generally fine to use without disclosure. The implicit standard here is that the candidate's claims about themselves are accurate, not that the prose is unassisted. Don't fabricate experience.

Scholarship and grant applications. Treat as you would a research paper. Most funding bodies have explicit AI policies as of 2025. Read them.

Penshift Chat is configured to decline if you ask it to write a complete piece for verbatim submission. Our human reviewers can suspend or terminate accounts where we observe clear patterns of misuse, including:

• Repeated requests for entire essays / dissertations / personal statements rather than refinement of submitted drafts.
• Attempts to fabricate citations or sources.
• Use patterns suggesting the account is being operated as a paper mill or essay-writing-for-hire service.
• Sharing the account with third parties contrary to our Terms.

Suspension is at our sole discretion. We may decline to refund accounts terminated for misconduct.

We don't recommend treating any single tool's output as proof of misconduct. Detectors disagree with each other constantly, and humanisation tools (including ours) shift the signal. The honest pedagogical move is to design assignments that AI alone can't complete well, ask for process artefacts (drafts, outlines, sources), and have a conversation when something looks off rather than relying on a percentage.

If you'd like to talk to us about institutional partnerships, classroom use, or research access, email [email protected].

We'll update this page as our position evolves. Material changes will be noted by a revision number above. Continued use of Penshift after a change means you accept the updated guidance.

See also our Terms of Service, Community Guidelines, and Privacy Policy.