Teaching & Learning
- If students are not told AI is allowed, they must assume it is prohibited for that assignment
- AI use in coursework varies by assessment category and by what the module leader or School authorises
- For Category 1 work, AI is prohibited; for Category 2 coursework, students may choose to use AI for support but the final submission must be their own words; for Category 3, AI use is required and assessed
In some modules and assessments, your module leader may permit the use of GenAI or even set tasks and assignments where using these tools is part of the challenge. Don’t use GenAI unless you have been told by your module leader that you can. If you’re unsure, check with them beforehand.
Category 1: Independent work
What this means for you: You must demonstrate your own knowledge and skills without GenAI assistance.
Category 2: AI-supported
What this means for you: It is completely your choice whether you use GenAI to support completion of your assessment. For example you could use GenAI to brainstorm, plan or refine, but the final submission must be in your own words and GenAI use is not assessed.
Category 3: AI-integrated
What this means for you: Demonstrating AI literacy is part of the grade.
Typical assessment formats: Same as Category 2 (e.g., reports, reflections, presentations, code notebooks), but with an added criterion that evaluates how well you use or critique GenAI and articulate its limitations, risks, and value.Purposeful GenAI use is required and assessed (e.g. quality of prompts, critical evaluation of outputs).
If the assignment brief says nothing about GenAI, assume its Category 1.
If no statement has been issued, or no response is forthcoming, then students should assume that the use of Generative AI tools to generate text, images or code (for example) is prohibited for that assignment.
- The university applies a three-category model to assessments
- The university expects Schools to tell students whether AI is allowed in each assessment, and if nothing is stated students should assume Category 1
- In-person exams, vivas, online tests, and practicals are typically Category 1, where AI may not be used; students in Category 1 assessments must work without AI assistance
Schools should provide their students with clear guidance on whether GAIT can be used in their assessments. The University has determined that there should be three categories of GAIT use for assessments:
• Category 1: GAIT may not be used
• Category 2: GAIT can be used to support student learning and development
• Category 3: GAIT use is actively encourage to help students develop their skills in the use of GAIT and understand how their use can be incorporated into assessment tasks
The University recognises that prohibiting the use of GAIT (i.e. under Category 1) will be difficult, but that there may be some areas where its use is inappropriate or not possible (e.g. in-person examinations, online tests, or practicals).
What this means for you: You must demonstrate your own knowledge and skills without GenAI assistance.
Typical assessment formats: Usually any assessment you do in-person, where its invigilated / supervised e.g. exams, vivas, practical lab tests
If the assignment brief says nothing about GenAI, assume its Category 1.
- The university still requires this use to be ethical and not to seek unfair advantage
- The university supports AI as a study aid and explicitly recognises its value for learning and employability
- Students may use AI to support learning and development in Category 2 contexts, and the annex gives examples such as summarising complex articles to aid understanding
The University is committed to taking a positive, educational approach to the use of GenAI as part of the teaching, learning and assessment activities you undertake. It aims to support the use of such tools – seeing their value as a useful aid to learning and their potential to enhance your educational experience and employability - while also maintaining the high academic standards and integrity of our assessments and awards.
GenAI must be used ethically and constructively, and in a way that doesn’t seek unfair advantage.
• Category 2: GAIT can be used to support student learning and development
It is assumed that the majority of assessments would fall under Category 2, where students are able to use GAIT to support them in exploring initial ideas and concepts, determining a structure for a piece of work, supporting their use of appropriate language, or otherwise support them in developing their ideas.
Learners with English as a second language or Specific Learning Difficulties might use the tools to summarise key points from complex articles, or otherwise aid their understanding.
- The university treats code-generating tools as GenAI and applies the same assessment-category rules to them
- If no permission is given, students must assume using AI to generate code is prohibited for that assignment
- Where AI is authorised, code-helper tools such as GitHub Copilot may be used within the relevant category rules
Code helpers (GitHub Copilot, CodeWhisperer)
If no statement has been issued, or no response is forthcoming, then students should assume that the use of Generative AI tools to generate text, images or code (for example) is prohibited for that assignment.
What this means for you: It is completely your choice whether you use GenAI to support completion of your assessment.
Typical assessment formats: Same as Category 2 (e.g., reports, reflections, presentations, code notebooks), but with an added criterion that evaluates how well you use or critique GenAI and articulate its limitations, risks, and value.Purposeful GenAI use is required and assessed (e.g. quality of prompts, critical evaluation of outputs).
Research
- The guidance also notes that publishers may impose their own AI rules for submitted papers and monographs
- AI use should be openly acknowledged, and material wholly or partly generated by AI should be clearly declared in the document
- For doctoral research, AI may be used for writing support such as improving grammar, structure, and clarity, but researchers remain responsible for the intellectual content and must not present AI-generated content as original work
Generative Artificial Intelligence (AI) tools—including Copilot, ChatGPT, and Gemini— offer new opportunities to support research, writing, and analysis.
You should also note that many publishers have their own guidance on the use of AI that you will need to comply with when submitting articles, papers or monographs during or after your doctoral research.
As a doctoral researcher, you are responsible for, amongst other things:
a) The intellectual content of your project.
AI tools are developing rapidly, as is their capability and reach. They can be a powerful research tool but their use should be openly acknowledged (see below).
By way of examples, doctoral researchers may use generative AI to:
• Improve grammar, structure, and clarity of their own writing
Researchers must not:
• Present AI-generated content as original work without proper acknowledgment
• Declaration of AI use. Material that is wholly or partially generated by an AI tool should be declared clearly in the document in which it occurs. Indeed, many publishers are now asking authors for a statement of AI use in their research papers.
- Uploading confidential, proprietary, or ethically sensitive data to public platforms is prohibited
- For data-sensitive tasks, researchers should use AI platforms approved by University IT Services, and the guidance states the university currently offers access to Copilot
- For doctoral researchers, AI may support research and analysis, but researchers must independently analyse their arguments, verify outputs, and must not fabricate data, sources, or citations
Generative Artificial Intelligence (AI) tools—including Copilot, ChatGPT, and Gemini— offer new opportunities to support research, writing, and analysis.
Researchers must not:
• Fabricate data, sources, or citations using AI tools
• Depend on AI to develop arguments without independent analysis
• Upload confidential, proprietary, or ethically sensitive data to public platforms
To ensure responsible use:
• Verify outputs - especially factual claims, citations, or code - for accuracy
• Use AI platforms approved by University IT Services for any data-sensitive tasks; the University currently offers access to Copilot.
• Protect personal and confidential information by avoiding uploads of unpublished material
- The university addresses research integrity and some funder-related AI rules, but does not define a unified university-wide policy for AI in grant proposals, IRB applications, or ethics declarations
- For externally funded projects, staff are told to check funder rules; the guidance says many funders expect disclosure of GenAI use in applications, do not permit GenAI in interviews, and that grant reviewers should not use generative AI in assessment activities
For externally funded Research and Innovation projects, you should consult your funder’s website and terms and conditions to check their policies on the use of GenAI during the application process. Many funders (including UKRI and the British Heart Foundation) expect researchers to disclose any use of GenAI on their application and do not permit applicants to use GenAI during interviews as part of the application process.
In addition, grant reviewers should not use generative AI as part of assessment activities. If you are unsure of the rules relating to specific funders, please contact your Research Engagement and Innovation Office (REIO) contact for further information.
Researchers thinking of using Generative AI to assist with funding applications should check if this is permitted in the funder’s policies. The Research, Engagement and Innovation Office (REIO) can be contacted with further questions on this.
Academic Integrity
- Students must acknowledge how and when they used each AI tool and make clear the extent of use
- Doctoral researchers are told that good practice includes identifying the user, software name and version, publisher, and date of use
- Disclosure is required when AI use has been authorised in student work, and doctoral guidance also requires clear declaration of AI-generated material in research documents
Key to the responsible use of Generative AI tools is the student’s intellectual ownership of the piece of work and the proper acknowledgement of the role played by the Generative AI tool in developing the piece of work.
Where Generative AI tools have been authorised by the School for a particular assignment, in order to maintain good academic practice and fairness in assessment, students must acknowledge when they have used Generative AI technologies in the development of their work. This includes acknowledging how and when they have used the Generative AI tools. Normally, such a statement should name each Generative AI tool used (providing a url for the tool) a description of each of the prompts or questions used, and all the outputs generated. It is each student’s responsibility to make it clear the extent and how a Generative AI tool has been used.
• Declaration of AI use. Material that is wholly or partially generated by an AI tool should be declared clearly in the document in which it occurs.
At the moment, there is no standard method to declare AI use but good practice includes citing the:
• User’s name
• Name and version of the AI software
• Publisher of the AI software
• Date the AI tool was used
- The provided sources do not define a university position on AI detection tools such as Turnitin or GPTZero
- Undisclosed or unauthorised AI misuse is treated as academic misconduct and would normally be considered plagiarism
- The university says suspected misuse will be investigated under normal academic misconduct procedures, and students are warned that failure to follow the rules can lead to disciplinary action
GenAI must be used ethically and constructively, and in a way that doesn’t seek unfair advantage. You should always abide by the guidance on Academic Integrity and Academic Misconduct, in the Assessment Handbook. Failure to do so could lead to disciplinary action.
The University recognises that there are legitimate uses of Generative AI tools while engaging with assessment. There will be assessments that direct students to engage with such tools in order to promote an understanding of how they work and to develop relevant skills and graduate attributes.
The use of Generative AI tools to prepare an assignment where there has been no specific instruction to do so, or where their use has been permitted by a School, or their use where not permitted (or beyond the scope of what was permitted by the School) would normally be considered as a form of plagiarism under the Academic Integrity and Academic Misconduct Policy (see section 9.2a(i) of Section 9 of the Assessment Handbook: Academic Integrity and Academic Misconduct).
Investigation and sanctions
Where a student is suspected of misconduct in their misuse of Generative AI tools their case would be considered under the usual procedures in section 9.5 of the Assessment Handbook: Academic Integrity and Academic Misconduct.
Institutional & Administrative
- Microsoft Copilot Chat is recommended for routine work use, but staff must follow privacy and data-protection limits
- The provided sources do not set specific university rules for grading, recommendation letters, or lesson planning beyond this general staff guidance
- The university supports staff use of AI in day-to-day work and provides specific staff guidance, while also encouraging staff to explore how AI can support students in assessment design
The University welcomes the introduction of GAIT in assessment design and encourages staff to explore how they might best support their students in understanding how to effectively use these tools, recognise their limitations, and understand both the ethical considerations and the advantages of using them.
This guidance aims to provide clarity on the University's expectations for staff when using Artificial Intelligence (AI) tools in day-to-day general working practice. It is in addition to the University's specific Data Protection and AI policies, and complementary to the use of GenAI tools guidance in teaching, learning and study.
The main pieces of advice in the Generative AI and AI meeting assistants guidance document are:
• The University acknowledges the benefits of AI and supports the use of AI tools.
• Microsoft Copilot Chat is recommended for day-to-day use at work. It can be used for tasks such as getting answers to queries, enhancing productivity and creativity, and DTS will provide basic support to colleagues with using it, as with our other Microsoft products.
• AI meeting assistants should not be used for work meetings without prior approval.
- The university has explicit data-protection restrictions for AI use
- Some tools are blocked or prohibited on university systems, including DeepSeek, and some meeting assistants require prior approval or must not be used
- Staff must not use personal, sensitive, confidential, or intellectual-property-related university or third-party data with generative AI models, and Microsoft Copilot Chat is the recommended approved platform for routine work, though it still must not be used for personal or sensitive data
It is essential that staff are aware of and apply this guidance when considering using such tools to ensure relevant University data is kept safe.
Personal or personal sensitive University data should not be used with Generative AI models. This means information related to an identified individual, and confidential information or intellectual property related to the University or third parties. If you have a particular need to use AI with data like this, please speak to DTS for advice.
Microsoft Copilot Chat is recommended for day-to-day use at work. It can be used for tasks such as getting answers to queries, enhancing productivity and creativity, and DTS will provide basic support to colleagues with using it, as with our other Microsoft products. Although Copilot has better data protection arrangements in place, it should still not be used for personal or personal sensitive data.
Supported: Digital Technology Services (DTS) can only provide technical help for Microsoft 365 Copilot Chat - this is freely available, and uses the same AI model as ChatGPT.
• Permitted but unsupported: You may use other GenAI tools, such as OpenAI ChatGPT, Google Gemini etc, and similar services - but DTS can’t offer any troubleshooting for them.
• Not permitted: DeepSeek is blocked and must not be used on University devices or networks.
AI meeting assistants should not be used for work meetings without prior approval.
Deep Seek Open source LLM Do not use Must not use – contact REIO and DTS for any research requirements
Teams Premium Paid for AI meeting capture assistant Supported – limited access Use on request and subject to use case approval
Otter.AI AI meeting capture assistant Not supported Do not use – request Teams Premium licence
Read.AI AI meeting capture assistant Not supported Do not use – request Teams Premium licence
- The university's stated strategy is supportive but aimed at preserving academic standards and integrity
- The university has an institutional governance structure and strategic approach for GenAI in teaching and learning
- It established a GenAI Working Group in August 2023, approved a university position statement through the University Board for Teaching, Learning and the Student Experience, and runs a GenAI Community of Practice to feed into strategy and good-practice development
Responding to the rapid uptake and widespread prevalence of Generative Artificial Intelligence (GenAI) tools circa November 2022, the University established a GenAI Working Group in August 2023 to consider its approach to the use of GenAI in teaching and learning.
The University's Position Statement on Generative Artificial Intelligence approved by the University Board for Teaching, Learning and the Student Experience (UBTLSE).
The University is committed to taking a positive, educational approach to the potential use of generative artificial intelligence tools (such as OpenAI’s ChatGPT, Perplexity, Google’s Bard, generative artificial intelligence (AI) assisted search engines and evolving functionality in Microsoft applications) as part of the teaching, learning and assessment activities undertaken by students.
The University aims to support the use of such tools, where appropriate, to enhance the educational experience and employability of students while at the same time maintaining the high academic standards and integrity of our assessments and awards.
The GenAI Community of Practice (CoP) is a multi-disciplinary forum for members of the University who are using GenAI for a variety of purposes – notably in teaching, learning and assessment.
The GenAI CoP is a discrete community whose outputs, advice and suggestions will be fed into the work of and considered by the GenAI Working Group, and also by CQSD, as part of the University’s strategy for Teaching and Learning (T&L).
DocuMark: Responsible AI Use for Academic Integrity
Knowing your institution's AI policy is step one. DocuMark helps enforce it fairly by empowering universities to manage AI-generated content, prevent cheating, and support student writing through responsible AI use.
Common Questions About University of Reading's AI Policies
University of Reading has defined AI policies in 12 of 12 categories, with an overall coverage score of 100%.
Disclosure is required when AI use has been authorised in student work, and doctoral guidance also requires clear declaration of AI-generated material in research documents. Students must acknowledge how and when they used each AI tool and make clear the extent of use. Doctoral researchers are told that good practice includes identifying the user, software name and version, publisher, and date of use.
Undisclosed or unauthorised AI misuse is treated as academic misconduct and would normally be considered plagiarism. The university says suspected misuse will be investigated under normal academic misconduct procedures, and students are warned that failure to follow the rules can lead to disciplinary action. The provided sources do not define a university position on AI detection tools such as Turnitin or GPTZero.
The university has explicit data-protection restrictions for AI use. Staff must not use personal, sensitive, confidential, or intellectual-property-related university or third-party data with generative AI models, and Microsoft Copilot Chat is the recommended approved platform for routine work, though it still must not be used for personal or sensitive data. Some tools are blocked or prohibited on university systems, including DeepSeek, and some meeting assistants require prior approval or must not be used.