Plagiarism and AI at King's: the declaration you must append

The short answer

King’s tells you exactly what to write. Every piece of coursework gets one of two published declarations appended with your references — one saying you used no AI, one saying you did and how.

It also stores your submitted work with your name attached, and publishes no way to get it out again. Both facts are worth knowing before you submit.

The two declarations

This is the most immediately usable thing on any page in this series. King’s publishes the wording:

“King’s requires students to acknowledge any use of generative AI tools in coursework by including a declaration statement along with your references. […] Unless alternative wording is suggested in an assessment brief then you should append one of the following:

1. I declare that no part of this submission has been generated by AI software. These are my own words.

2. I declare that parts of this submission has contributions from AI software and that it aligns with acceptable use as specified as part of the assignment brief/ guidance and is consistent with good academic practice. The content can still be considered as my own words.”

(The grammatical slip in the second statement is in King’s own text.)

The second declaration comes with a tick-list. You acknowledge use of software to: “(i) Generate ideas or structure suggestions, for assistance with understanding core concepts, or other substantial foundational and preparatory activity”; “(ii) Write, rewrite, rephrase and/or paraphrase part of this essay”; or “(iii) Generate some other aspect of the submitted assessment” — each with the tools and links inserted.

And the reassurance students most need, stated by King’s itself: “so long as acknowledged use falls within the scope of appropriate use as defined in the assessment brief/guidance then this will not have any direct impact on the grades awarded.” Declaring honestly does not cost you marks.

Note that the assessment brief can substitute different wording — “unless alternative wording is suggested in an assessment brief”. Check there first.

Separately, the Policy requires a general own-work declaration on everything: students must “make a declaration with each assessment to confirm that the assignment submitted is their own work.”

Grammar checking still counts as no AI

This single note answers a question that worries a lot of students, particularly those writing in a second language:

“Note. Using software for English grammar and spell checking is consistent with Statement 1.”

Statement 1 is the one that says no part of the submission was generated by AI. So running a spellchecker or a grammar checker over your own sentences does not force you into the AI declaration.

That is a sensible, liveable line and it matches how most universities in this series draw it. It is also where our own check is designed to sit — it marks filler, hedging and tired transitions in your own sentences and suggests plainer wording, and nothing is rewritten for you. But read the next section before assuming any tool is covered, because King’s is precise about where assistance stops.

Where the help has to stop

Clause 4.3 is one of the most exactly drawn boundaries in the series, and it applies equally to a human proofreader, a private tutor and a piece of software:

“Where a student chooses to ask another individual to proof read their work, engages a private tutor to assist with assessments, or uses writing technologies, this should be limited to learning activities which do not form part of the final stages of assessments submitted for credit, or highlighting errors in spelling, punctuation or grammar. The service should not assist with substantive content creation or structuring of the assessment. Above all, third parties cannot make substantive changes which compromise the authorship of the text (for example, translating whole sections of text, rather than individual phrases), revise calculations, formulae, equations, or code, alter charts, figures or diagrams, correct information or references, or involve any tutoring.”

Four items in that list catch people who were not trying to cheat:

• Structuring. Having something reorganise your argument is outside the boundary, not merely editing it.
• Translating whole sections. Individual phrases are fine; passages are not.
• Correcting information or references. If a tool tells you a fact or a citation is wrong and you change it, that is outside the permitted help.
• Tutoring. A chatbot explaining your topic during the final stages of an assessment is tutoring.

And the general principle in 4.2: “the use of generative AI tools to produce output which is then copied as part of a submission is not appropriate… any work submitted must represent a genuine demonstration of the student’s own work, skills and subject knowledge.”

AI is a third party, not a new offence

King’s did not create an AI offence. It deemed AI a third-party service and used the machinery it already had:

“For the avoidance of doubt, generative Artificial Intelligence writing programmes are third party technologies and must only be used in accordance with College guidance.”

The two categories it lands in are “third party involvement” — “when a student receives unauthorized assistance from a third party which results in the submission of an individual assessment which the College cannot be satisfied wholly represents the student’s own work or understanding” — and “third party production”, where the work “is substantively the product of a third party”.

One consequence of that design choice: King’s Academic Glossary, which the policy says holds the definitions of misconduct types, carries no entry for artificial intelligence at all. If you go looking for one, that is why.

What King’s stores about you

This clause deserves reading carefully, because it is the sharpest privacy contrast in this series:

“Any material presented for assessment may be submitted to an academic misconduct detection service for text analysis […] Submitted work will be stored in a database (along with the student’s name, email address, programme/module details and institution) and will form part of the body of student work against which future submissions from this and other institutions will be compared.”

Your name goes in with the work, and the comparison pool extends to other institutions. No opt-out, withdrawal or deletion route is published anywhere in the policy or procedure.

Compare Cambridge, which bars names outright — “student names will not be used, including in the filename” — and grants a right to remove your work from the database after it has been checked. Two UK universities, opposite answers.

The practical implication is one worth applying anywhere: before you put a draft into any service, find out what it keeps and what it attaches to it. For what it is worth, our own check does not add your text to any third-party comparison database.

The eleven things a decision-maker looks at

King’s publishes no statement about whether Turnitin’s AI-writing detection is switched on, so we will not tell you. What it does publish is the full evidence list an Assessment Sub-Board Chair considers, and a detection report is one line of eleven:

“The student’s submission • instruction, guidance or requirements for the assessment task • a similarity or authorship report produced by TurnitinUK or other relevant software • search engine results linked to the submission content • original source material identified • academic judgement of relevant markers/module leads • Examination Invigilator’s Reports • Any statement or evidence provided by the student • Notes of any relevant meeting or discussion […] • Evidence from any other relevant parties where relevant • Any history of academic misconduct by the student.”

Two things to take from that list. A report sits alongside academic judgement and original source material, not above them. And your own statement and evidence is an enumerated item — so providing drafts, version history and an explanation is not pleading, it is submitting evidence the decision-maker is required to consider. See how to prove you wrote it yourself and what to do when you are accused.

If a detector score is what started this, our guide to an AI detector flagging your thesis sets out what that number does and does not establish on its own.

The word “authorship” in that clause is notable — it is the term Turnitin uses for its authorship products — but King’s does not name the feature or say whether it is enabled. On what such reports can support, see what a Turnitin score actually means and how accurate AI detectors really are. Our own published error rates are measured on English and German corpora, and no detector produces proof.

Sanctions

Where a Misconduct Committee finds a charge substantiated on the balance of probabilities, the published restorative outcomes include: referral to support services; a formal warning with the unaffected parts marked and the result capped at the pass mark; assignment of the minimum pass mark; cancellation of the results with a mark of zero; and conditions for the continuation of student status.

Note the second of those — a warning combined with marking only what was unaffected, capped at a pass. For a first, contained case that is a relatively survivable outcome, and it is worth knowing it exists on the published list.

Cases are administered by Student Conduct & Appeals within the Student and Academic Services Directorate. The Academic Misconduct Policy took effect on 1 September 2023, superseding the previous Misconduct Policy, the Academic Honesty and Integrity Policy and the Proof Reading Policy, with a stated next review of September 2026 — so check whether a newer version has appeared.

Guidance does not beat policy

One structural point you should know, because it affects how much weight to put on the declaration wording above.

The AI guidance — first published October 2023, last updated April 2025 — says of itself:

“This is guidance not policy […] It should be considered alongside the Academic Policy and Procedure […] If the reader finds any apparent inconsistency between this Guidance and the Academic Policy and Procedure or any other policy then please do note the precedence of Policy over guidance. Failure to do so could lead to misconduct proceedings being taken.”

So the declaration wording is published by King’s, is what King’s asks for, and is the right thing to use — but it sits in a subordinate document, and the binding Academic Misconduct Policy does not itself prescribe an AI declaration form. If your assessment brief says something different, the brief and the policy come first.

Before you submit

• Append one of the two declarations with your references, unless your brief gives different wording.
• Spell and grammar checking is consistent with declaring no AI— King’s says so.
• Do not let anything restructure your argument, correct your facts or references, or translate whole sections.
• Assume your name is stored with your work. There is no published opt-out.
• Keep drafts and version history. Your own evidence is an enumerated item in the decision.

If you want a read on your own text before you hand it in, our AI check is free up to 1,500 words, with no account and no name.

Sources

• King’s College London: Academic Misconduct Policy — Approving Authority: Academic Board; effective 1 September 2023; next review September 2026.
• King’s College London: Generative AI guidance for students — first published October 2023, last updated April 2025; the declaration wording. Expressly guidance, not policy.
• King’s College London: Academic Glossary — the misconduct-type definitions; carries no entry for artificial intelligence.

This page is orientation, not legal advice. What binds you is your assessment brief and the Academic Misconduct Policy.