The explosion in the use of AI, and in particular Generative AI (GenAI), has implications for copyright both in terms of the data used to train tools such as Large Language Models (LLMs), and in the outputs from these tools (text, images, multimedia).
This is a fast-moving area, and the lack of harmonisation of copyright laws globally is a further complicating factor. Some of the issues are being addressed via legal cases (particularly in the US), and by legislation, namely the EU’s Artificial Intelligence Act. Within the UK, discussions about how best to balance the rights of copyright holders whilst encouraging AI driven innovation are ongoing but have yet to produce specific guidance.
Inputs and prompts
When entering prompts, or inputting text or other material into GenAI tools, check terms and conditions to see whether your input/s will be used to train the tool.
- If the answer is yes and the material that you are inputting is your own work, consider whether you are happy for it to be used in this way. For example, if you enter original work, your ideas will become part of the “commons” without, and before, you have been recognised for your work.
- If the material you are inputting is not your own work, do you have the copyright holder’s permission to upload it or is it in the public domain or licensed in a way that permits it being used in this way? (For example, although CC BY licensed material allows re-use and remixing, it requires attribution of the original source.) Please consider your use carefully, particularly if the GenAI tool uses inputs as training data.
If you have any questions about this, please contact the Copyright Officer in the first instance. Information on AI and Text & Data Mining can be found on the TDM page.
Outputs
Please be aware that there is some debate about whether most GenAI outputs qualify for copyright protection under UK law, given that for a work to qualify for copyright protection it must demonstrate the author’s own intellectual creation. Until the legal position becomes clearer, we suggest that you consider the points below.
Before re-using outputs:
- Read the Terms & Conditions of the AI tool that you’re using to check who owns any rights that exist in outputs. If it is not the person who entered the prompt, check how you are permitted to re-use the content.
- Consider whether the text or image (for example) substantially reproduces original content. Could the output be viewed as a derivative work or adaptation of an existing work? If so, by sharing the output you could be infringing copyright.
LSE guidance
The School’s Research Guidance and AI Legal and Regulatory Guidance [forthcoming] will include information on copyright.
Generative AI: Developing your AI Literacy (LSE staff and students only) This Moodle course provides a useful introduction to Generative AI, covering areas such as how it works, how to use it ethically and how it is being used in industry. Sitting within section 2.2: Risks and Harms of Generative AI is a short section entitled ‘Disputed Ownership and Copyright Infringement’ which we encourage you to read.
Further reading
JISC have published guides on various aspects of Generative AI. An introduction to copyright law and practice in education, and the concerns arising in the context of GenerativeAI by John Kelly summarises the issues and current uncertainty around copyright and GenAI.