UnTXT and University College London's AI Policy - What Students Need to Know

Summary: UCL's approach to generative AI is "supportive rather than prohibitive." Using UnTXT to audit your own work aligns with UCL's expectation that your submission is "substantially your own work." Detection tools are part of transparent scholarly practice.

What UCL's AI Policy Actually Says

UCL ties the question to a single word, and the word is creating.

Its guidance asks you to acknowledge generative AI "in any piece of academic work where it has been used as a functional tool to assist in the process of creating the academic work." It then says acknowledgement may not be needed for things like "helping to improve your grammar and writing structure" — the tidying-up end, rather than the making end.

UnTXT sits further from creating than a grammar checker does. It writes nothing, suggests nothing and changes nothing. It reads a finished draft and marks the passages that carry machine-writing patterns, and then it stops.

What UCL does not do is give you a blanket answer, and we are not going to pretend otherwise. Its guidance is explicit that "it is the nature of the task that determines whether use needs acknowledgement or not. This will be discipline and assessment dependent, so it is for teaching staff to determine." Where it is unclear, UCL's own instruction is plain: "Students should ask teaching staff if further clarification is needed."

So the honest answer for a UCL student is that your department decides, and asking is cheap.

What we would do. Mention it anyway. One line in your submission — "I used UnTXT, an AI detection tool, to check my own draft" — takes a sentence and closes the question permanently. Declaring something you did not strictly need to declare has never gone badly for anybody.

Source: Acknowledging and referencing GenAI — UCL Library Skills, checked 7 September 2026. UCL also publishes assessment-level guidance through its Generative AI Hub. Policies change; where UCL's pages and this one disagree, theirs is the one that counts.

UCL's Distinctive Approach

Unlike some universities that ban AI entirely, UCL recognises that AI is a tool many students use. Rather than prohibit it, UCL's policy asks you to:

Use it responsibly - Not to replace your thinking, but to support it
Disclose it - Be transparent about what tools you used and how
Maintain authorship - Your final submission should be substantially your own work
Show your thinking - In papers and assignments, demonstrate your reasoning, not just a polished output

UnTXT supports all four of these expectations. It's not a tool AI forbids; it's a tool the UCL framework actively supports.

Why UCL Students Use UnTXT Before Submitting

You've used AI for brainstorming or drafting—but you've rewritten it

This is the most common scenario at UCL. The policy allows it, as long as the final submission is substantially yours. But how do you know the rewrite is thorough enough? UnTXT shows you which passages still carry AI patterning, so you can address them before submission rather than guessing.

You want to understand what "substantially your own work" means in practice

UCL's language is deliberately flexible, and "substantially your own" is open to interpretation. UnTXT will not resolve that for you, and it would be dishonest of us to suggest it can. What it gives you is a prompt: a grey passage is one worth rereading and asking whether it sounds like you. Sometimes the answer is that it does, and the passage stays exactly as it is. A mark is a question, not a finding.

You're preparing for an academic integrity review

If your work is questioned, having run UnTXT and revised based on the results is protective evidence. It shows you took the integrity standard seriously and audited your own work.

How to Use UnTXT for a UCL Assessment

Finish your draft - Complete your work including any AI-assisted brainstorming or editing
Read the grey passages first - grey means the passage reads as machine-patterned; gold means it reads as yours
Re-analyse - Run UnTXT again to confirm flags are resolved
Acknowledge in your statement - Add a line to your assessment disclosure: "I used UnTXT to audit my draft for AI patterns"

Why this workflow aligns with UCL's policy:

You're being transparent about your process
You're ensuring substantial personal authorship
You're demonstrating good scholarly practice
You're using detection as an audit tool, not an evasion tool

What the Results Mean for Your Submission

UnTXT's flags are signals, not verdicts. A flagged passage doesn't mean you used AI—it means that passage has statistical characteristics common in AI-generated writing.

At UCL, where the policy allows thoughtful AI use, a flagged passage might simply mean:

You used formal academic register (normal for essays)
You quoted or paraphrased a source (and didn't acknowledge it)
You drafted with AI and didn't fully rewrite
Your writing style differs from earlier work (FORENSIC signal)

The modules most relevant to UCL assessments:

Module	What it flags	Interpretation

Frequently Asked Questions

Can I use AI in my UCL assessment?

According to UCL's policy: yes, you can use it, but your submission must be substantially your own work and you must disclose your use transparently. Check your department's specific guidance for any additional constraints.

If I used AI, do I have to tell UCL?

Yes. Transparency is central to UCL's framework. If you used AI (even just for brainstorming), acknowledge it in your assessment statement. Hiding it is worse than disclosing it.

What does "substantially your own work" actually mean?

It means most of the intellectual work, reasoning, and writing is yours. You can use AI as a tool (like you'd use a spell-checker or grammar tool), but the final submission should reflect your thinking and voice.

If UnTXT flags my essay, will UCL flag it too?

Not necessarily. UnTXT and UCL's review process use different approaches. But if UnTXT flags large sections, it's worth revising so you're confident in the work's authenticity.

Can I check my essay with UnTXT and not mention it?

Technically yes, but best practice is to acknowledge it. It demonstrates you took the integrity standard seriously. Most teachers respect transparency.

How much does it cost?

The first 1,000 words are free on sign-up, no card needed. A typical essay runs 2,000–3,000 words, so that allowance covers roughly the first third of one. Checking whole pieces, or checking regularly, means a monthly subscription from £8.40 including tax, which carries a monthly CVT allowance.

UCL's Philosophy & Why It Matters

UCL's approach is notable in the UK higher education landscape. Rather than treating AI as a threat to academic integrity, the university treats it as a feature of modern scholarship that students need to navigate thoughtfully.

This philosophy assumes:

Students are intelligent enough to use tools responsibly
Transparency is more important than restriction
Detection tools support learning, not just enforcement
Your job as a student is to develop your thinking, not to avoid tools

UnTXT aligns with this philosophy. It's not a surveillance tool; it's a learning support tool that helps you ensure your work genuinely reflects your own thinking.

If You're Reviewed for Academic Integrity

If your work is questioned:

Be honest about your process - Explain exactly how you used AI, if at all
Provide evidence - Drafts, notes, earlier versions showing your thinking
Run UnTXT yourself - Before the meeting, analyse your work. If you've already done this and revised based on it, you have a strong defence
Reference UCL's policy - You were operating within the policy framework

Most cases resolve through conversation. Your goal is to provide clarity and evidence that you acted in good faith within UCL's framework.

The Bottom Line

UCL's AI policy trusts you to use tools responsibly and to be transparent about it. Using UnTXT to audit your work before submission is exactly what that trust looks like in practice.

Start for free - first 1,000 words. No payment needed. Sign up at untxt.uk

This page reflects UnTXT's understanding of UCL's publicly available AI policy as of early 2026. Policy details may change. Always check UCL's official guidance and consult your department's specific requirements.