FAQs on GenAI tools

As part of its ongoing digital transformation, the University is partnering with OpenAI to provide access to ChatGPT Edu for all students and staff (including College staff), from the start of the new academic year (2025/26). All staff and students will receive an activation email to access their individual ChatGPT Edu account. Training and support are also available, across the University. Further information can be found on the GenAI at Oxford website.

This is part of the University’s ongoing collaboration with OpenAI and its commitment to ensuring safe and responsible use of GenAI tools and embracing the opportunities they provide for enhancing teaching and learning, research, and administration at Oxford. 

OpenAI is a US-based company which has developed some of today’s leading generative AI products, such as ChatGPT (including ChatGPT Edu) and previously, DALL-E. OpenAI is also a Microsoft partner and you can read the latest news on this on the OpenAI website. 

ChatGPT Edu is an AI tool that has been specifically designed for use in higher education, offering enhanced privacy, security, and OpenAI’s flagship GPT-5 model. The University's ChatGPT Edu workspace is a unique ChatGPT environment, with its own settings, members, and resources. These ensure that data is retained within the University.
 

Generative AI (also known as GenAI for short) is a type of Artificial Intelligence with a chatbot interface, that can be used to create original content such as text, images, video or audio, in response to prompts. This content is based on patterns learned from vast amounts of data the GenAI tool has been trained on.

Artificial Intelligence (AI) is a form of computing that can perform human-like tasks such as problem solving, decision-making, and pattern recognition. Generative AI is a specific type of AI that can produce new content. Machine learning refers to AI technology that can recognise patterns in data.

The University has strategic partnerships with both OpenAI and Microsoft. Since 2024, the University has been working in strategic partnership with Microsoft to enhance its research, education, and digital infrastructure. This includes provision of Microsoft’s GenAI tool, Copilot Chat, as part of the University’s Microsoft 365 package.

The University’s collaboration with OpenAI announced in March 2025 focuses on expanding our AI offering and capabilities. OpenAI is also a Microsoft partner (Copilot is built on OpenAI’s GPT-5 technology) but ChatGPT and Copilot use this technology in different ways.

You can purchase licences through the AI Competency Centre for enhanced Copilot functionality (Microsoft 365 Copilot) for £26 per user, per month. Find out more (including key use cases for ChatGPT Edu and Microsoft 365 Copilot) on the AI Competency Centre website.

There has been a yearlong pilot of ChatGPT Edu involving around 750 participants, including academics, research staff, postgraduate research students and professional services staff, in a wide range of roles across the University and Colleges. Students have also been involved in a series of funded projects through the AI Teaching and Learning Exploratory Fund with the Centre for Teaching and Learning, to explore how AI can be used to enhance education at Oxford. 

This decision was taken by Information and Digital Committee (IDC) in Trinity term 2025, following the results of a successful pilot and informed by discussions at Education Committee, who felt there was a need for equitable access to GenAI tools for all students, and Research & Innovation Committee, who recognise the use of AI tools in research and have developed guidance on its use.  

Following an examination of options, the IDC agreed that the offer available through the University's collaboration with OpenAI was the most suitable and offered best value for money in relation to the University's requirements.  

The Committee will continue to examine the range of options available for secure access to generative AI, in light of changing technology and our evolving understanding of the way the technology is being utilised across the institution, and to examine other tools and platforms through managed pilots, where appropriate, to ensure that provision meets business need effectively.

IDC is chaired by the Pro-Vice-Chancellor (Digital) and includes representation from across the divisions and colleges, including the chairs of the IDC sub-committees. These chairs include the Pro-Vice Chancellors for Education and Research, the Head of Gardens Libraries and Museums and the Registrar.

ChatGPT Edu accounts are being made available free at point of use, across the University and Colleges, as part of an inclusive offer. The rollout is funded by the Digital Transformation programme – a multi-year, strategic programme aiming to enable the University to continuously evolve its digital services and activity to meet the challenges and harness the opportunities of the digital era.

Yes. The Bodleian is delivering a pilot digitisation project with OpenAI in 2025, as part of the University's wider partnership. This project includes work to digitise out-of-copyright materials held in the Bodleian’s collections, to help scale up a programme of ongoing digitisation and to explore the potential for generative AI to help. Findings of an initial pilot are due to be compiled later this year and shared by publishing research outputs on the University of Oxford’s research repository, ORA, so that the wider community can benefit from lessons learnt.

You can find out more about this work in the March 2025 announcement about the University’s expanding collaboration with OpenAI.

To support the rollout of ChatGPT Edu across the University and Oxford colleges, our AI training programme has been expanded. All students can access free, ongoing training and support, delivered by in-house specialists in IT Services, at the AI Competency Centre, Bodleian Libraries, Oxford Lifelong Learning, and the Oxford Research Software Engineers Group.

To help you get started with ChatGPT Edu, you can book online training or in-person training (SSO required) through the Digital Capabilities team in IT Services. Our workshops and courses emphasise the importance of human involvement, critical thinking, and ethical and responsible application.

Recordings of key sessions will be shared on Canvas during Michaelmas term 2025 (SSO login required). You can also access free training in getting started with ChatGPT Edu, any time, through the OpenAI Academy.

Many people across the University have been using ChatGPT linked to a University department/College or SSO account. The University is now providing inclusive access to secure ChatGPT Edu accounts for all staff and students. ChatGPT Edu has been designed specifically for the higher education sector, with enhanced data privacy and security.

If you have one of these pre-existing accounts, look out for an email from the University’s ChatGPT support team with information about what to expect and any action you may need to take. This will depend on how you currently access your ChatGPT account (e.g. whether you use your SSO credentials or your department/college email address).

If your (non-Edu) ChatGPT account is linked to your University SSO account, it is important to be aware that you will no longer have access to your previous chats once your account moves to the University’s secure Edu workspace, unless you follow the prompt when you log in to merge your chat history, including any Custom GPTs you may have created (recommended). You will also have the option to export and delete your old data.

If you have a ChatGPT account linked to your departmental email address you will now also have access to the University’s secure ChatGPT Edu, which is linked to your SSO. To reduce the data risks of using your departmental email linked account for work or study purposes, you are advised to close it and open another free ChatGPT account if you wish to, using a private email address instead. ChatGPT accounts not associated with the Oxford (Edu) workspace should only be used for non-work or study activity. They are not secure or approved for processing confidential or personal data and should not be used for any process involving University, staff or student data. You should start using your University-supported ChatGPT Edu account linked to your SSO account for all work or study-related activity. Your SSO-linked ChatGPT Edu account will also stay with you if you move colleges, departments or change email address.

If you have paid for a ChatGPT licence (e.g. Plus or Pro) other than ChatGPT Edu or ChatGPT Team/Business, for an account linked to your University email or SSO, you will receive a refund from OpenAI for any remaining time on that licence. You should use your University ChatGPT Edu licence for any work or study-related activity in future. The OpenAI refund will happen automatically unless you have a mobile subscription. Mobile subscriptions must be cancelled in your app store and you will receive instructions for how to do that. Licences for ChatGPT Business (formerly Team) accounts will not be refunded, as in this case, you will be able to access the University’s Edu workspace and continue your work there. If you experience any issues or would like to discuss specific details of your licence, please contact the IT Service Desk; if you need to discuss refunding, please use this feedback form.
 

Yes, but it is important to consider the security implications of each tool. The University is rolling out ChatGPT Edu as part of an expanding provision of secure, enterprise-level generative AI tools. However, this does not limit your use to one specific tool. Individuals have different preferences, and different AI tools may be better suited to specific tasks.

It is important to ensure any AI tool you are using is secure. Secure, University-supported tools include Microsoft Copilot when accessed via SSO, and Google NotebookLM and Gemini. Visit the AI Competency Centre for more information on using AI tools securely, or speak to the team for support and guidance.

For guidance on data compliance when processing data using generative AI tools, please refer to the University’s mandatory Information Security and Data Protection training. For specific queries about use of data in this context, contact the Information Security team in the Assurance Directorate for advice.

All students with a Single Sign-On (SSO) account should have to received an activation email from OpenAI with details of how to access an individual ChatGPT Edu account. If you have not received an email by the end of Freshers' Week (10 October), please complete this online form to request support. Please also check your junk mail folder first, in case the email is there.

You can also log in to ChatGPT through chatgpt.com. To learn more about accessing your new ChatGPT Edu account, please refer to this ChatGPT Edu Onboarding Guide from the University’s AI Competency Centre. This will also take you through ChatGPT’s interface and help you get started. It is important that you log into your account using your full University Single Sign-On (SSO) username in the exact format [email protected].

Once your Single Sign-On (SSO) ceases to be active, your access to the University’s Edu workspace will also end. You will need to export what you need from your account before you leave, subject to the relevant data policies.

It is not possible to automatically merge or migrate these accounts, but to preserve useful information you can take the following steps:

• For account memory, open Settings → Personalization → Memory → Manage, and copy any memories saved in this section.
• For custom instructions, open Settings → Personalization → Custom Instructions in your old account and copy anything you have added (e.g., how you would like ChatGPT to respond, or context it should remember). Paste this into the same section of your new Edu SSO account to keep a consistent experience.
• To make sure you can still access custom GPTs publicly published in the Oxford workspace you use regularly, go to GPTs → Explore, search for the GPTs you rely on, copy the URL, and start a new chat in your SSO-linked Edu account with each, to save them in your sidebar.
• To transfer any personal GPTs you have created that are not publicly published, go to GPTs → Explore → My GPTs and recreate them in your Edu workspace. Alternatively, if the information within the GPT can be made public, you may choose to publish it to others at Oxford University. With this sharing setting enabled, you will then be able to access your GPTs from your SSO account. Guidance for sharing custom GPTs can be found on the AI Competency Centre website.

Only SSO-linked ChatGPT Edu accounts should be used for work purposes, as these benefit from multi factor authentication login and offer enhanced privacy and security. Your more secure, SSO-linked account is private and will remain with you if you move between departments or colleges within the University, unlike an account linked to your department or college email address.

No. Migration of content into the Edu workspace is one‑way and irreversible, so you will not be able to export this data later (e.g. if you subsequently leave the University).

No, personal subscriptions accessed via personal emails are not impacted. If you have set up a personal ChatGPT account using University credentials (e.g. department or SSO login), however, you will receive an email letting you know how your account is affected, if any action is required, and you will also find information about this in the FAQs above. 

This happens very rarely. It essentially means you are given a new University identity. In this case, you should coordinate with the IT Service Desk in advance to discuss migration impact and options for your ChatGPT Edu account and access to other services that are likely to be affected.

Visit the GenAI tool features page to view details on the models and features you can access. 

No, ChatGPT Edu is provided for University-related work only. Under the University's IT Regulations, computer facilities and networks may be used solely for purposes directly connected with the work of the University and its colleges, and with the normal academic activities of their members. 

There are ongoing debates about the implications of generative AI – from concerns about its social and environmental impacts, to its application within academia and effects on human cognition.

Use of AI is a complex and rapidly developing arena. Teams across the University are involved in a wide range of research programmes on AI, including its application in education and its societal implications, and impacts on health and wellbeing. You can find out more about some of this research on the AI at Oxford website published in 2023.

The Oxford Martin School is currently inviting Expressions of Interest for a solutions-focused research programme with joint funding from OpenAI (as part of its NextGenAI initiative) and the Oxford Martin School, to explore impacts of the development and widespread use of generative AI. Selection and delivery of the successful programme will be entirely led by the Oxford Martin School.

Thousands of members of staff and students are already using this technology in their work and study at Oxford and within this context it is important the University provides a secure environment for those who choose to use the technology and explore the opportunities it can provide – supported by training and guidance that encourages safe and responsible use.

The University’s guidance and policies in relation to GenAI are under continuous review, and many departments/faculties and colleges have developed, or are in the process of considering guidance which relates to their specific context and academic field of enquiry.

Use of generative AI for study is a matter of personal choice. Both staff and students are advised to follow any advice from their tutors and supervisors, departments, faculties or colleges, as regards appropriate use within specific academic disciplines and courses. Use of GenAI by students should always be in consultation with course providers.

The rapid rise in global adoption of AI is associated with significant demands on energy and water consumption. At Oxford, AI is relatively small contributor to the University’s overall digital footprint, but as its adoption grows, the resulting emissions are an important consideration.

At the same time, the technology itself is rapidly developing, bringing improvements in the energy consumption of AI models, hardware, and data centres, and innovations in areas such as water recycling and heat redistribution from data centres.

Inclusive access to secure AI platforms is equipping Oxford’s researchers to pursue technological advances and breakthroughs in energy and climate science. Recent work enhanced by GenAI to tackle the significant emissions associated with MRI/CT scans is one example of this work in practice (read more in this case-study on Regent Lee’s research in the Medical Sciences Division).

AI’s energy consumption comes in two parts: the initial energy required for training new models, which is usually a large amount over a short period of time; and the energy cost per individual query, which is relatively low per prompt, but can add up over time. The University’s agreements with AI companies mean that data processed in University-provided tools are not used to further train AI models.

It is challenging to get a real understanding of the true energy consumption and environmental implications of GenAI from tech companies: a lot depends on the application and model being used, and available approximations vary considerably.

What we do know is that, in comparison to many common, everyday tasks, AI's carbon footprint is relatively small. It is significantly lower than driving or showering, for example, or video conferencing, but higher than simpler digital activities such as web browsing. Google estimates that one average AI message in Gemini chat is equivalent to about 9 seconds of watching TV.

The University is actively engaging with tech companies such as OpenAI, Google, and Microsoft to push for more sustainable practices. Oxford’s collaborations mean it can directly question providers on how they are delivering on their net zero targets.

Several AI vendors have published statements on their green technology ambitions, such as Microsoft’s commitment to becoming carbon negative by 2030 and Google’s ambitions for more sustainable and efficient operations as part of its 2030 net-zero ambitions. Google reduced its overall data centre emissions by 12% in 2024, compared to 2023, and it recently published research into the energy consumption of its AI tools here.

OpenAI has published a commitment to minimising its environmental impact and reducing energy consumption at all stages of AI development. On its sustainability ambitions, an OpenAI spokesperson said: “We’re focused on making our AI more capable and more efficient. Each generation of our models uses less energy to train and run, and our new Norway data centre will be powered entirely with renewable energy. We’re also teaming up with researchers to explore how AI can speed up progress on clean energy and climate science.”

You can read more about Generative AI and carbon emissions in the Oxford context specifically in this article on the Oxford Students website.

The environmental impact of GenAI is the subject of ongoing conversations across the University. There are research teams dedicated to exploring this – such as iDoddle, a project looking at the impacts of digitalised daily life on climate change within the Environmental Change Institute. A new £1 million Oxford Martin School research programme has also been announced, in partnership with OpenAI, to explore solutions for applying gen AI to global challenges or exploring its impact. The Centre for Digital Scholarship (Bodleian Libraries) and GLAM Environmental Sustainability also held an event on The Environmental Impact of Digital in Michaelmas term 2025.

FAQ updated 19 December 2025

The University’s Environmental Sustainability Strategy was approved in March 2021 and its net-zero ambitions are reflected in University decision making. As part of our governance processes, for example, committee papers are required to outline sustainability impacts.  

Given the rapid global adoption of AI since 2021, it is becoming increasingly important to examine how people are using these technologies, to understand the impact on our environmental targets and to be able to make purposeful steps to address that impact.

The University’s inclusive provision of secure GenAI tools is helping us to understand more about how people at Oxford use AI and to encourage greater transparency and more open conversations about its use. The more we understand about how people are using AI, the better equipped we will be to develop appropriate responses to mitigate the environmental impact.

The Sustainability Strategy is reviewed in five-year cycles, and the first five-year review is already well underway. While emissions from GenAI are a relatively small part of the University’s overall digital footprint − let alone the environmental footprint of the institution as a whole − addressing the increasing impact of AI adoption is a key consideration for our next five-year cycle.

A great deal of work is being done, in the meantime, to assess emissions associated with the University’s third-party providers (including AI data centres). This will help us to make better-informed and purposeful decisions about mitigating this impact to support our sustainability ambitions. Work is also underway to explore guidance for University members on how to limit the environmental impact of using AI and other digital services for work and study.

Opportunities are also being explored to harness GenAI technologies to help make our research and other activity more efficient, and to reduce emissions elsewhere in the University. Smart use of GenAI to improve the way we monitor and control the heating and lighting of buildings (given most of our impact relates to our estate) could also accelerate a reduction in our overall environmental impact.

Outcomes of the Environmental Sustainability Strategy review will be shared with colleagues during the 2025/26 academic year.

FAQ updated 19 December 2025

ChatGPT Edu is a secure generative AI platform, which means there is no data retention beyond the University and ChatGPT models cannot be trained on data from within our ChatGPT Edu workspace. This is not the case with freely available ChatGPT accounts.

ChatGPT Edu is an AI tool specifically designed for use in higher education, offering enhanced privacy and security. To support the safe and secure use of AI, all staff are reminded to refresh their mandatory Information Security and Data Protection training, which must be taken each year. This now includes a section on use of AI, with advice on processing data in AI tools. Find out more about using generative AI tools safely. Find out more about handling personal data.

No, not in normal circumstances, your account will be private and accessible only to you, in a similar way to how you use your University email account. However, the principles of the IT Regulations that relate to Examining User’s Data will apply to the contents of your prompts/outputs (as with other IT tools and services), so in the case of certain investigations you may be required to present a record of your prompts and outputs.

Your ChatGPT data cannot be seen by OpenAI, and it won’t be used to train AI models. In certain, limited circumstances (for example, if the system detects that someone has expressed a plan to harm others during their use of ChatGPT), there are a restricted number of people at OpenAI who can potentially review specific content to escalate concerns if they deem there is a risk of physical harm. As AI usage becomes more widespread, we will continue to review our strategy and guidance on privacy and access.

The University takes student welfare very seriously, with robust wellbeing support services in place. Your college and department offer support with welfare and wellbeing. You can access mental health and wellbeing support from Student Welfare and Support Services (SWSS). There is also information and advice in the Welfare and Support section of the Student Handbook 2025/26.

The University provides all students with access to Togetherall, a free and anonymous online platform where students can connect with others about what’s going on in life - big or small. Togetherall includes a range of free courses, resources, and anonymous chats to support your wellbeing, available 24/7.

You can also use the free counselling service from SWSS, and you will find a wealth of useful resources online, via the Oxford Students website or your MyOxford app. The counselling team can refer you to specialist mental health services or to the University’s Disability Advisory Service which can advise you on study and assessment related adjustments and support.

In addition to the central Student Welfare and Support Services, you can get advice and support from your college, department, fellow students, or the Students' Union, and there are contacts you can use in both emergency (including out of hours) and non-emergency circumstances, for health, welfare and academic support.

Generative AI tools can be used to support and enhance work, teaching and learning, but they are not a replacement for human interaction and are not equipped to offer therapy or mental health support. GenAI tools lack emotional nuance and clinical awareness, and it is important to be aware that they can give inaccurate and potentially dangerous advice.

At times of heightened distress or mental health crisis, AI chatbots may reinforce harmful thoughts through feedback loops, which are self-reinforcing cycles where by negative thoughts, emotions, and behaviours reinforce each other, often leading to a downward spiral in mental wellbeing (e.g. when anxiety about a job interview or exam leads to poor sleep, which in turn makes you feel unprepared, confirming your fears, and reinforcing the cycle).

The University continues to monitor developments in GenAI and to review and update guidance on safe and responsible usage by staff and students. Training is now available to all staff and students on safe and responsible use of Generative AI tools such as ChatGPT.

Find out more about how OpenAI is improving its models to support wellbeing, including expanding interventions to help people who may be in crisis and making it easier for them to reach emergency services and get help from experts.

AI tools are improving all the time, but they do make mistakes (e.g., through ‘hallucinations’) and they have biases. ChatGPT Edu is a powerful tool, but content should always be human-checked and referenced. Generative AI should always be considered a tool to support rather than a replacement for human review and decision making.

No, you do not have to use ChatGPT Edu or any other AI tools unless elements of your course or assessment specifically require their use. We are rolling out access to ensure inclusive access for all students and to provide access to technology you are likely to encounter in future, in an AI-enabled world. Whether you choose to use these tools is up to you but if you do, you should use them in line with University guidance on safe and responsible use. 

The University-wide rollout of ChatGPT Edu provides a secure GenAI environment for study, supported by training and guidance which emphasises ethical usage, the importance of critical thinking, and responsible application in line with University policies. Guidance on the safe and responsible use of these tools has been updated for this academic year.

This guidance includes information about use of GenAI in materials submitted for assessment during your time at Oxford. For each specific assessment you will be given prior written notice of whether the use of AI is permitted. You are fully responsible for the accuracy, originality and quality of work you submit for assessment, and you must follow the rules on AI use for each specific assessment you are taking.

Cases of suspected unauthorised use of AI in assessment are covered by the University’s student conduct regulations. You should make yourself familiar with these regulations, as any actions that are found to be in potential breach will be reviewed in line with the University’s Code of Discipline and Academic Disciplinary procedure.

You should always follow the guidance of your tutors, supervisors, and your department or faculty.

Teaching staff may use AI to improve efficiencies in the course of their work. For example, they may use a GenAI tool to speed up administrative work or improve processes (by automating routine tasks, for example), in order to free up more time for creative thinking and direct support for students. 

The Oxford teaching model requires you to engage critically with complex ideas, to compare and contrast ideas, and to demonstrate a thorough engagement with the academic literature, through face-to-face learning, supervision and tutorials.

GenAI tools like ChatGPT Edu have the potential to further personalise, enrich, and enhance your study. The University provides guidance on using GenAI safely and responsibly, to develop your academic skills.

The University is supporting work to explore the opportunities that AI provides, including in education. Last year, the Digital Transformation Programme funded technical and pedagogic support to help staff bring some of their ideas for this to life. Building on outcomes of this work, you can expect to see AI features appearing in Canvas and other digital education tools.

No, any research collaborations with OpenAI are subject to the same regulations and scrutiny as any other research partnership with an external organisation. Later in the academic year, there will be a call for proposals for a new research programme on AI funded by OpenAI and the Oxford Martin School. While OpenAI has provided funding to support this academic study, through its ongoing collaboration with the University, it has no role in the selection of the research programme as well as no further role in the research programme itself, in terms of any editorial control, involvement in the research activity, or influence on outcomes before publication. Through research partnerships such as the one with the Oxford Martin School and OpenAI, and other examples in different parts of the University, we hope to deepen our understanding of the usage and implications of AI, to inform our evolving approach.

For help and support with your GenAI tools:

1. Consider first seeking support from the company whose product you're using:

• ChatGPT Edu: For access and activation queries, please use the OpenAI help portal. You can use the chat at the bottom of the page to ask questions and connect with support by clicking on the Message tab. If you are unable to resolve your issue through the help portal, then you should contact OpenAI support directly by email at [email protected].
• Microsoft Copilot: The Microsoft Support site can be searched to assist with queries on Copilot.
  Google Gemini: Google's Gemini Apps Help page offers various pieces of advice and support for using Gemini.
• NotebookLM: Google’s NotebookLM Help page covers help topics from getting started to troubleshooting NotebookLM.
• If your query is about whether an outage is affecting your connection to a particular tool, try the relevant status pages:

- OpenAI Status page for ChatGPT Edu

- Microsoft Status page for Microsoft Copilot

- Google Status page for Google Gemini and NotebookLM

2. For simple queries regarding University-supported GenAI tools and support, use the AI Competency Centre chatbot to be directed to relevant information.

3. For support accessing AI services, contact the Central IT Service Desk - they can help you with ChatGPT Edu, Microsoft Copilot, or the GenAI tools in the University's Google workspace (Gemini and Notebook LM). If your query needs escalating, the IT Service Desk will involve our GenAI specialists at the AI Competency Centre or signpost you to tailored support from elsewhere in the University, as needed.

Further support and inspiration is available via our growing network of AI Ambassadors, and through the Generative AI Special Interest Group (SIG) on Microsoft Teams, which everyone is welcome to join. This forum is a great place for conversations about developments in Generative AI and sharing ideas.