Generative AI tools analyse text, images, code and other material for machine-learning purposes, retaining information and creating new content in response to prompts provided by users.

Material may have become part of the AI model’s training dataset as a result of web crawling, licensing agreements with content providers such as publishers, or user-upload files and interactions (prompts). This raises new and complex considerations in terms of how the tools we use intersect with the rights of authors and creators, and how AI can be used safely and ethically in the context of teaching, learning and research activities. Considerations will depend on the tool(s) and material used, and the specific activity and context in each case. 

The University’s recommended AI tools are Google Gemini, which can be used to generate, analyse and summarise text as well as help you find information, and NotebookLM, an AI powered assistant that helps with research and note-taking. Note that these tools function differently in terms of how they interact with copyright material; Gemini works with a range of sources in its training dataset which may or may not be cited, whereas NotebookLM is designed for users to provide uploaded or linked sources which it then works with.

There are still many contested and unresolved challenges and questions, so current University guidance advises caution when it comes to using copyright material in the context of generative AI tools. This is an area of rapid change, and it’s beneficial for users to stay informed about developments around how these tools work and emerging issues in relation to copyright. We will endeavour to keep this guidance up to date, but it is not comprehensive in scope and should not be interpreted as legal advice. 

There are also important, related concerns around using AI tools with data which is sensitive, confidential, personal or subject to specific licensing agreements. These considerations are not addressed specifically in this Practical Guide, but are covered by the University guidance linked below. 

It is important to be aware of the following University guidance, all of which emphasises the risks involved when using generative AI tools and copyright material.

You should read the guidance relevant to your own work and context for examples of acceptable and unacceptable uses of generative AI:

• IT Services general guidance on generative AI tools (Dec 2024): "You must not upload any copyrighted content, including journals, to any GenAI tools unless you have confirmed this is acceptable with the publisher". 
• Generative AI - Staff guidance and Taught students guidance (2024): “Avoid uploading personal/sensitive/copyrighted material to generative AI”
• Artificial Intelligence use in assessment - Student guidance (Nov 2023): “Warnings: Data security, intellectual property and ethics: providing your own work to an individual or software has a degree of risk. Once you have shared your work, you cannot guarantee how that will then be used. The emergence of generative AI, for example, has raised serious concerns about how data is processed and used by the companies and software.”
• Guidance on the use of generative AI in PGR programmes (2024): “Your research must meet legal and University expectations in terms of… intellectual property rights (including copyright)”; “Generative AI… may produce outputs that ignore intellectual property rights, for example by producing content that does not include appropriate acknowledgement or breaches copyright.” 
• Responsible AI Use in Research: Policy & Best Practice Document for Researchers (May 2025): “Research, especially unpublished work and third-party content, must not be shared with AI systems without assurances regarding data protection, copyright, intellectual property and without anonymisation.”

The University has also signed up to the Russell Group principles on generative AI in education (Feb 2024), which outlines our role in supporting students and staff to become AI-literate. This includes understanding the opportunities, limitations and ethical issues around plagiarism and copyright infringement. 

Jisc has issued guidance on Generative AI and copyright law and practice in education (Mar 2024) which raises some 'challenging questions' as a starting point for navigating copyright issues:

There is currently no legal framework in the UK around copyright and generative AI specifically. However, there are broad legal exceptions which may be applied when using copyright material in the context of private study, non-commercial research or as part of an assessment or teaching activity.

These exceptions are explored in further detail elsewhere in this Practical Guide (see Copyright law explained) and depend on the following considerations:

• You must only copy material which is relevant to your research, study or teaching activity (consider using a short quotation or extract rather than a full article, for example);
• The material must not be used to train a publicly-accessible AI tool or transferred to any other users, including contacts at other institutions or the general public;
• You must have lawful access to the material used (e.g. through a Library subscription), and must not circumvent any technical protection measures or restrictions put in place by a rights holder or content provider (this is specifically in relation to text and data mining processes, which may require users to obtain a licence or access to an API or platform from the provider);
• In the case of research, this must be for a non-commercial purpose (be mindful of activities which take place in collaboration with commercial partners, or with commercialisation in mind);
• Sufficient acknowledgement must be provided, where possible and necessary.

If these assurances cannot be met, then you should avoid uploading copyright material unless you have specific permission to do so from the relevant rights holder. 

Note that UK legislation is more restrictive in terms of ‘fair dealing’ exceptions than countries such as the US, where concepts such as ‘transformative use’ may be applied to a broader set of purposes. This is significant in terms of the global scale of generative AI development, which is the basis upon which regulatory changes have been proposed to encourage innovation whilst protecting the interests of rights holders (see Gov.uk: Copyright and Artificial Intelligence consultation (Dec 2024)). These proposals have proved controversial, however, with concerns summarised by the likes of CREATe (the Centre for Regulation of the Creative Economy) in their consultation response working paper (Feb 2025). 

Any eventual changes or new guidance issued by the Government will hopefully provide clarity in this area, and this Practical Guide will be updated as and when appropriate. 

Generally speaking, the risk of infringement will be reduced when using openly-licensed materials such as open access journal articles and publications which have been made available under a Creative Commons licence with reduced restrictions on sharing and reuse.

It is important to note that open licenses operate within the legal framework of copyright and their own terms of use. For example, a basic requirement in all Creative Commons licences is that attribution must be provided, and in some cases there are restrictions on commercial use or onward sharing of derivative (modified) versions of the material, which could include AI-generated content.

There is also a lower risk when using public domain materials, which is where copyright protection has expired or has been waived by the rights holder. It is still considered good practice to attribute the authors of public domain materials, where possible.

Generative AI tools are emerging which have been developed entirely using openly-licensed and public domain materials (e.g. Common Pile v0.1), and initiatives are taking place to increase the availability of out-of-copyright works for use as training material (e.g. Harvard Institutional Data Initiative). 

For further guidance on Creative Commons licensing and public domain tools, see Creative Commons for Researchers: a Practical Guide and the For Research page (Using copyright material in your research).