Accused of Using AI in Your Assignment? What to Do Next

Receiving an email about suspected unauthorised AI use can be concerning, particularly when the allegation relates to work that you believe reflects your own research and writing. The appropriate response depends on the circumstances and your university's rules, so the first step is to identify the specific concern being raised rather than assuming that every AI-related allegation follows the same process.

This guide explains what a Turnitin AI-writing result can and cannot establish, what types of academic evidence may help provide context about your work, how university procedures can differ, and where students can seek further guidance. The aim is to help you respond accurately and responsibly rather than to provide methods for changing your writing to influence an AI detector.

If you have legitimate records of your academic work, such as earlier drafts, research notes, source material or document version history, keep them in their original form. These records may help you explain the development of your work if your university asks you to discuss the submission. Only provide material that is relevant and appropriate under your institution's procedure.

Understanding an Academic Misconduct Allegation in the UK

Receiving an email about suspected unauthorised AI use does not, by itself, establish that academic misconduct has occurred. The university's correspondence should explain what concern has been raised, which assessment or submission it relates to, and what procedure you are being asked to follow. The exact process and terminology can differ between universities and sometimes between types of assessment.

It is also important not to assume that every AI-related concern is automatically treated as the most serious form of academic misconduct. Some institutions distinguish between different types or levels of academic-integrity concern, while others use their own terminology and procedures. Your university's regulations are therefore more important than a general description of how UK institutions handle these cases.

In particular, the Skills and Post-16 Education Act 2022 should not be confused with a university's internal disciplinary process. The Act created offences relating to the provision, arrangement and advertising of commercial contract-cheating services in England. It does not establish a universal university procedure for investigating suspected AI use, nor does it determine whether an individual student's submission amounts to academic misconduct.

Similarly, there is no single UK-wide rule that every university must use the same evidential standard or impose the same penalty for an AI-related allegation. Universities operate under their own academic regulations and procedures, which can specify how concerns are investigated, how students can respond, and what outcomes may be available.

What a Turnitin AI-Writing Report Can—and Cannot—Show

A Turnitin AI-writing result should be treated as an indicator for further review, not as conclusive proof of who wrote a piece of work. Turnitin states that its AI-writing detection may misidentify human-written, AI-generated and AI-paraphrased text. It also advises that an AI-writing result should not be used as the sole basis for adverse action against a student.

This distinction matters because an AI-writing report does not independently establish the circumstances in which an assignment was produced. A percentage or highlighted passage cannot, by itself, explain whether a student wrote the work without AI assistance, used an AI tool in a way permitted by the assessment rules, or breached a specific university requirement. Those questions require the relevant assessment instructions, university regulations and the surrounding evidence to be considered.

Why an AI-Writing Result Should Be Interpreted Carefully

AI-detection systems make predictions from patterns in submitted text. That means their output has limitations and should be interpreted in context. Turnitin itself describes its AI-writing detection as not always accurate and recommends further scrutiny and human judgement when considering whether academic misconduct may have occurred.

It is therefore not appropriate to treat particular writing characteristics—such as formal academic language, consistent paragraph structure, sophisticated vocabulary or the use of conventional signposting—as proof that AI generated the work. Students are often taught to write clearly, structure an argument logically and use academic conventions. Those characteristics can occur in genuinely human-written assignments.

What You Should Not Do After Receiving an AI Flag

If your work has been flagged, avoid changing your writing simply to try to produce a different detector result. Deliberately adding awkward wording, varying sentence lengths for the purpose of affecting detection, inserting irrelevant personal details or using AI-paraphrasing tools does not establish authorship and may create additional academic-integrity concerns.

Instead, concentrate on information that genuinely explains how the assignment was researched and developed. Depending on your university's procedure, relevant material may include:

• Earlier drafts: Keep legitimate versions of the assignment that show how the work developed over time.
• Research notes and source material: Retain notes, reading records and sources that are relevant to the work.
• Document version history: Where available, genuine version history can help provide context about the development of a document.
• Permitted AI-use records: If your assessment allowed specific forms of AI assistance, retain relevant records and follow the disclosure requirements in the assessment guidance.
• Assessment instructions: Keep the original guidance showing what forms of AI use, editing or other assistance were permitted for that particular assessment.

The key question is not how to make your writing look less detectable. It is whether you can accurately explain how the work was produced and demonstrate that your conduct complied with the rules that applied to the assessment.

If you believe an AI-writing result has been interpreted incorrectly, address the specific allegation through the university's stated procedure. Focus on factual information about your work and the applicable assessment rules rather than attempting to reverse-engineer or manipulate the detection system.

Gathering Legitimate Evidence of Your Work

If your university asks you to respond to an AI-use allegation, gather genuine records that can help explain how the assignment was researched, developed and submitted. The purpose is not to manufacture a defence or create evidence after the event. Instead, provide relevant material that already exists and follow the university's instructions about what evidence it will consider.

There is also no single document that automatically proves who wrote an assignment. The usefulness of a particular record depends on the circumstances, the assessment requirements and the university's procedure. A clear and consistent collection of legitimate materials can, however, provide useful context for the work under review.

Drafts and Version History

If you genuinely developed your assignment using a document platform that retains earlier versions, preserve those records in their original form. This may include Microsoft Word or Google Docs version history, depending on the software and account you used.

1. Keep the original document: Do not overwrite or reconstruct earlier versions simply to create a more convincing timeline.
2. Preserve relevant versions: Where version history is available, retain earlier drafts that genuinely formed part of your writing process.
3. Show meaningful development: Earlier versions may help explain changes to your argument, structure, references and wording.
4. Follow the university's instructions: Submit version-history material only where it is relevant and permitted under the procedure you have been given.

Do not assume that document metadata establishes the exact amount of time you spent writing or proves that every word was produced without assistance. Different applications record different information, and metadata can have limitations. The strongest approach is to present authentic records accurately rather than relying on a technical field as definitive proof.

Research Notes, Sources and Earlier Work

Your research process may also provide useful context. Keep genuine notes and source material that relate directly to the assignment, particularly where they help explain how you developed the ideas, evidence or references used in the submitted work.

• Research notes: Keep genuine handwritten or digital notes used during your research and planning.
• Source material: Retain relevant journal articles, books, reports and other sources that you actually consulted.
• Annotated materials: Where appropriate, preserve legitimate annotations, reading notes or highlighted passages that show your engagement with sources.
• Earlier drafts: Keep outlines, partial drafts and revisions that genuinely formed part of the assignment's development.
• Assessment guidance: Retain the assignment brief and any instructions explaining whether AI tools or other forms of assistance were permitted.
• Permitted AI use: If AI assistance was allowed for the assessment, retain relevant records and disclosures where the university requires them.

Browser history can sometimes provide background about online research, but it should not be treated as definitive proof of authorship. Only provide it if the university's procedure requests or permits it, and avoid submitting large amounts of unrelated personal browsing information.

Most importantly, do not describe academic support, proofreading or other assistance as evidence of independent authorship unless that description accurately reflects what happened and the assistance was permitted under the assessment rules.

What to Expect at the Academic Misconduct Panel

If your initial written response does not resolve the AI accusation, you will be summoned to an academic misconduct panel. While the thought of defending yourself before faculty members is intimidating, understanding the structure of these hearings will help you present your evidence calmly and effectively.

An academic misconduct panel in the UK is an internal university procedure, not a legal trial. It typically consists of a neutral chair (often a senior academic from a different department), faculty members, and a note-taker. The objective of the panel is to review the evidence submitted by both you and the module leader to determine, on the balance of probabilities, whether academic misconduct occurred.

The Role of Your Students’ Union Representative

Your SU representative is an independent advocate who understands the specific university regulations better than you do. They will not speak for you—you must present your own case—but they provide critical support by:

• Ensuring the panel strictly follows the university’s published procedural rules.
• Intervening if the questioning becomes overly aggressive or strays from the specific allegation.
• Helping you organize your evidence, such as your version history and research notes, into a coherent presentation.

How to Present Your Evidence Calmly

When you are called to speak, avoid emotional appeals or accusations against the software. Focus entirely on the facts.

• Acknowledge the Flag, Pivot to the Process: Begin by calmly stating, "I understand why the software flagged my work, but I am here to demonstrate the manual process I used to write this assignment."
• Walk Through the Timeline: Present your compiled digital footprint chronologically. Show the panel your initial research notes, library checkouts, and the exact metadata timestamps from your early drafts.
• Address the Text: If the panel points to a specific paragraph that Turnitin flagged as 100% AI, pull up your version history. Show them the exact keystrokes: "You can see here on the 12th of November at 4:00 PM, I typed the first half of this sentence, paused to consult my notes, and then completed it ten minutes later."
• Offer a Live Demonstration (If Necessary): Offer to log into your Google Docs or Microsoft 365 account live during the hearing to prove the metadata screenshots are authentic.

Rebuilding Confidence: How to Avoid Future AI Flags

Surviving an academic misconduct panel is exhausting, and it is entirely normal to feel anxious about submitting your next assignment. However, you can immunise your future work against algorithmic false positives by deliberately adjusting your writing style to reflect messy, highly specific human thought.

1. Master Genuine Lexical Cohesion

Stop relying on robotic transition words like "Furthermore," "Moreover," or "In summary" to start your paragraphs. Sandra Matthias, our Lead Global Editor for Humanities, advises students to focus on conceptual links rather than stock vocabulary:

"True academic writing isn't about using complex words; it's about achieving perfect lexical cohesion where every paragraph effortlessly advances your core thesis."

When you link paragraphs by carrying a specific argument from the end of one section directly into the beginning of the next, you naturally lower the text's predictability (known as perplexity). This effectively bypasses the primary trigger for AI flags.

2. Vary Your Sentence Structures

Algorithmic text generators inherently favour uniform sentence lengths. To prove human authorship within the text itself, intentionally vary your rhythm. Follow a long, complex sentence analysing a piece of academic literature with a short, direct, declarative statement summarising its specific impact. This variance (known as burstiness) is a hallmark of human writing.

3. Insert Highly Granular UK Case Studies

The ultimate defence against automated detection is hyper-specificity. Generalised, sweeping statements trigger high AI probability scores. If you are writing a legal essay, do not simply summarise the broad concept of corporate governance; integrate the precise jurisdictional nuances found in a UK law assignment guide, citing exact sub-sections of recent English judicial precedent. If you are studying nursing, reference a specific, localised NHS Trust operational policy rather than making broad observations about "the healthcare sector."

By anchoring your essays in granular, localised evidence, you create a distinctly human draft that algorithms cannot reliably mimic.

If you are struggling to structure your next assignment and need zero-AI academic mentorship or structural blueprints to guide your research safely, connect with our academic mentors at the Global Student Advisory Desk.