Your law dissertation deadline is close, ChatGPT or a similar tool could plausibly save you hours on structure and prose — and you genuinely do not know whether using it crosses a line your department will treat as misconduct. Law has one risk no other undergraduate subject shares as sharply: AI tools have been documented fabricating case citations and legal authorities that do not exist. Here is what is actually allowed, how to disclose it, and the workflow that keeps you safe.
What Do UK Universities Actually Say About AI Use?
There is no single UK-wide rule — each university sets its own academic integrity policy, and QAA (the Quality Assurance Agency for Higher Education) has published sector guidance rather than a binding rule, including “Maintaining quality and standards in the ChatGPT era” and “Reconsidering assessment for the ChatGPT era” (2023), which focus on redesigning assessment and promoting academic integrity rather than mandating a single disclosure format across the sector. What this means practically: read your own law school’s specific AI-use policy before assuming a rule from another department, another university, or a general article applies to you. Where a policy allows AI assistance with structure, editing or explanation but requires the legal analysis, argument and final prose judgement to be your own, that distinction is the one to hold onto regardless of the exact wording your department uses.
What Is AI Genuinely Useful for in a Law Dissertation?
- Structuring an argument you have already researched. Turning a set of case notes and your own analysis into a logical chapter outline is a legitimate editorial task.
- Explaining a concept you have already read primary sources on. Using AI to check your own understanding of a case’s reasoning, after reading the judgment yourself, is different from using it to skip reading the judgment.
- Improving clarity and flow in your own draft. Editing for readability is a widely accepted use, similar in kind to a proofreading tool.
- Generating a first-pass OSCOLA footnote format for a citation you have already verified is correct. The formatting mechanics, not the substance of what is cited.
What Is the One Risk Specific to Law?
AI tools generating fabricated case names, citations, or even entire quotations attributed to real judges — commonly called “hallucinated” case law — is a documented, real risk, not a hypothetical one. Because legal argument depends entirely on the accuracy of the authorities cited, this is a categorically different risk from a fabricated statistic in a business dissertation: a fabricated case in a law dissertation is a fabricated primary source, and submitting one, even unknowingly, is a serious academic integrity issue as well as a substantive legal error. The rule that follows from this is absolute and non-negotiable: verify every single case name, citation and quotation an AI tool suggests against a primary legal database (Westlaw, LexisNexis, or the free BAILII) before it appears anywhere in your dissertation. Never trust an AI-generated citation on its face, however plausible it looks.
A Worked Example of the Verification Step
Illustrative: an AI tool suggests that a named 2019 Court of Appeal case established a particular principle about contractual interpretation, complete with a plausible-looking neutral citation. Before that sentence goes anywhere near your dissertation, the check is: search BAILII or Westlaw for the exact case name and citation given, confirm the case actually exists, open the judgment itself, and confirm the principle the AI attributed to it is actually what the court held — not a paraphrase that has drifted from the original reasoning. If any one of those four checks fails, the citation does not go in, regardless of how confident or specific the AI’s original suggestion sounded. A citation that “sounds right” is not evidence that it is right.
A Safe Workflow for Using AI in a Law Dissertation
- Do your primary reading first. Read the actual cases, statutes and academic commentary yourself before any AI involvement — AI assistance should come after your own understanding, not instead of it.
- Never ask AI to generate case citations from memory. If you need to find authority for a point, search a primary database (Westlaw, LexisNexis, BAILII) yourself rather than asking an AI tool to supply one.
- Use AI only on your own already-researched material for structure and clarity. Feed it your own notes and draft, not a request to “find cases about X.”
- Verify every citation that appears anywhere near AI-assisted text. Even if you wrote the surrounding prose yourself, re-check every case name and pinpoint citation against the primary source before submission.
- Disclose according to your department’s exact policy. Where your department requires a statement of AI use, follow its specific format precisely rather than a generic acknowledgement.
- Keep a record of your primary-source reading. A simple reading log naming which cases and statutes you read directly, and when, is useful evidence of genuine engagement if your process is ever questioned.
How Do I Disclose AI Use Correctly?
Where AI assistance touches your dissertation, both an explicit acknowledgement statement (if your department requires one) and, in some cases, an in-text reference are expected — these are two different obligations, not one. Our guide to referencing ChatGPT and generative AI in Harvard style covers the published Harvard template and the separate acknowledgement wording UK departments commonly expect, and is directly applicable here regardless of your specific subject.
How Does This Interact With Your Broader Research Process?
If AI assistance touches your literature review specifically — summarising academic commentary, structuring your reading — the same fabrication risk applies to secondary sources, not just cases: an AI tool can invent a plausible-sounding journal article or author that does not exist. Our guide to writing a dissertation literature review with AI, honestly covers the broader verification discipline this requires across any subject, which the case-law-specific rule above sits alongside rather than replaces. And where your dissertation is doctrinal in structure, our guide to what a doctrinal law dissertation is and how to structure one covers the primary-vs-secondary source discipline a doctrinal methodology already demands — the AI verification habit above is really an extension of that same discipline, not a new one. If you are still at the proposal stage, our guide to writing a law dissertation research proposal is worth reading alongside this one, since a proposal that states your intended AI-use approach up front tends to avoid the ambiguity that causes problems later.
What If a Supervisor or Marker Asks Me About My Process?
Being able to explain, specifically, which parts of your dissertation involved AI assistance and how you verified any resulting citations is a reasonable expectation, not an accusation — treat a supervisor’s question about your process the way you would treat any methodology question, with a direct, specific answer rather than a defensive one. Keeping the reading log mentioned above makes this straightforward: you can point to exactly when you read the primary source that underlies a given citation, independent of any AI involvement in the surrounding prose.
What Mistakes Cost the Most Marks — or Worse — Here?
- Submitting an AI-suggested citation without verifying it. This is the single most serious risk in this guide — a fabricated case is both a substantive error and a potential academic integrity issue.
- Assuming your department’s policy matches a friend’s at a different university. AI-use policies vary; read your own department’s current policy directly.
- Treating “disclosure” and “referencing” as the same obligation. Many departments expect both, separately — an acknowledgement statement and, where the AI output itself is referenced, a proper citation.
- Using AI to generate legal analysis rather than to structure your own. Where a policy distinguishes editorial assistance from substantive analysis, the line is usually whether the legal reasoning is genuinely yours.
- Assuming OSCOLA formatting help means the underlying citation is correct. Formatting assistance and substantive verification are two different tasks — do both.
- Having no record of your own primary-source reading. If your process is ever questioned, a simple reading log is far stronger evidence than your memory of what you read and when.
Once your research and citations are verified, Tesify drafts your dissertation chapters from your own verified research and argument, free to start, and every word stays yours. Tesify never invents a case, statute or source — it structures and drafts around the material and citations you provide.
Frequently Asked Questions
Can I use ChatGPT or similar tools for my law dissertation at all?
Check your own department’s specific policy first — most UK law schools allow some editorial and structural AI assistance while requiring the legal analysis and argument to be genuinely your own, but the exact line varies by institution.
Do I have to disclose AI use in my law dissertation?
Increasingly expected, and often required by department policy — follow your own department’s exact disclosure format, which may include both an acknowledgement statement and, where relevant, an in-text reference.
Can AI really invent fake case law?
Yes — this is a documented, real phenomenon, sometimes called “hallucinated” citations. Never include an AI-suggested case, citation or quotation without independently verifying it against a primary legal database first.
Is using Tesify the same as using ChatGPT for my dissertation?
Tesify is built to draft and structure around your own verified research and citations rather than generate legal claims or case law itself — check your department’s policy either way, and always independently verify any legal citation regardless of the tool used.
What is the safest way to use AI for my literature review?
Read your primary and secondary sources yourself first, use AI only on your own notes and drafts for structure and clarity, and independently verify any author, article or citation AI output references before it appears in your dissertation.
Will using AI for editing get flagged as plagiarism?
Editing assistance for clarity is generally treated differently from generating original analysis, but similarity-detection tools and academic integrity processes vary — follow your department’s specific guidance rather than assuming either way.
Does the QAA have a single national rule on AI use in dissertations?
No — QAA has published sector-wide advisory guidance on assessment design and academic integrity in the AI era, but individual UK universities and departments set their own specific AI-use policies.
What should I do if I am not sure whether my AI use is allowed?
Ask your supervisor or your department’s academic integrity office directly before you submit — this is faster and safer than guessing from a generic article, including this one.
How do I keep a reading log for primary sources?
A simple spreadsheet or document noting the case or statute name, the date you read it, and a one-line note on what it established is enough — the point is having a record independent of your memory, not a formal system.