> ## Content Index > Fetch the complete content index at: https://www.peridot.works/llms.txt > Use this file to discover other available public pages before exploring further. # UCL coursework and AI: check permission before planning the workflow - URL: https://www.peridot.works/ucl-ai-assessment-brief-permission/ - Published: 2026-10-19T02:42:22.000Z - Updated: 2026-10-19T02:42:22.000Z - Description: For UCL coursework, distinguish tool availability, assessment-specific permission, the stated assistive-use default and accurate acknowledgement. - Author: Santosh Srinivas An AI tool being available to students does not settle how it may be used in a particular assessment. For a UCL master's student, a useful starting point is the actual assessment brief: what work is being assessed, what assistance is permitted and how any use should be acknowledged. This guide uses UCL's student guidance and Generative AI Hub checked on 19 October 2026\. It concerns coursework, not application documents, and does not replace module instructions. UCL says its assessment categories are being reviewed during 2026/27; that notice is not itself a new permission or prohibition. ## Separate the tool from the task UCL's [student assessment guidance](https://www.ucl.ac.uk/study/current-students/exams-and-assessments/assessment-success-guide/engaging-generative-ai-your-education-and-assessment?ref=peridot.works) says teachers explain whether and how GenAI may be used. If instructions are unclear, students should ask the module leader. Where no explicit instructions are given, its stated default is an assistive role, with the final submission still substantially the student's own work. Keep that institutional default distinct from any more specific assessment instruction. Do not borrow permission from another module, another student or a demonstration of a tool. Similarly, a tool's ability to generate a plausible answer does not establish that using the answer is appropriate for the work you are being assessed on. Read the brief alongside the relevant handbook or module guidance. Note the version and the assessment to which it applies. If two instructions seem inconsistent, identify the exact conflict when asking for clarification rather than choosing the more convenient interpretation. ## Describe the assistance you are considering Make a small planning record with three fields: the learning task, the proposed assistance and the instruction that governs it. Be precise about the operation. Checking a sentence you wrote, replacing whole passages, generating an analysis and reviewing your own explanation are different activities. For an original hypothetical example, a student may want feedback on whether their explanation is clear. The useful question is what feedback is permitted and how they will decide what to change themselves. This is a planning example, not permission to upload an assessment or a claim that UCL approves every feedback tool. Do not describe extensive rewriting as merely checking spelling. Record the actual scope of assistance. A workflow that changes during the task should prompt another check of the applicable instructions, not an unchanged label in the final acknowledgement. ## Acknowledgement does not create permission UCL's student guidance calls for acknowledging GenAI assistance. It also says presenting an entire AI-written assessment as your own is unacceptable. These are related checks: permitted use still needs the required transparency, and an acknowledgement does not authorize use that exceeds the brief. Follow the current departmental and library guidance for the acknowledgement required by your task. Keep an accurate private process note so you can explain what you did, what assistance you used, what you rejected and what you changed yourself. That note is an editorial suggestion for honest preparation, not a mandatory UCL submission template. Avoid inventing a tool version or claiming you verified information that you did not inspect. If a reference or factual claim enters your work, locate and evaluate the original evidence rather than treating a generated citation as proof that the source exists or supports the claim. ## Availability and assessment permission remain separate The [UCL Generative AI Hub](https://www.ucl.ac.uk/teaching-learning/generative-ai-hub?ref=peridot.works) links student resources and identifies institutionally provided tools and support. The hub's access information and its notice of the category review should be read alongside the actual assessment guidance, not treated as a blanket rule for every task. Use the university's current advice when deciding what information is appropriate to enter into a service. Do not experiment with confidential placement material or someone else's work simply to see whether a tool responds. A fictional or synthetic practice example can help you understand a tool without exposing real records. ## Keep learning and explanation in view Before submitting, check whether you can explain the reasoning, sources and choices in the work you are presenting. If an unfamiliar explanation has appeared in the draft, investigate it and decide whether it belongs there. A polished sentence is not evidence that the underlying reasoning is sound. The [assessment and workload guide](https://www.peridot.works/masters-assessment-workload/) helps identify what a course actually asks students to produce. The [application-AI policy guide](https://www.peridot.works/masters-application-ai-policy-check/) addresses a different document context; admissions rules should not be imported into coursework or vice versa. Your next action is one precise question if needed: for this named assessment and proposed operation, what assistance is permitted and what acknowledgement is required? Keep the answer with the relevant instructions. The aim is a workflow you can explain honestly, with the learning and authorship the assessment requires still your own.