Gen AI and Copyright Checklist: Putting it in Practice

• GenAI tools are trained on large amounts of data that are most likely protected by copyright. Unless they are using the data with permission, under a licence or under a copyright exception, AI tools could be infringing copyright.

• Do not input any materials or information that is private or commercially sensitive.

• Do not input any third-party copyright materials to GenAI tools, or social media platforms unless you have written permission or they are licensed in a way that allows sharing. Third-party material could mean photos, illustrations, songs, videos, data, text. If you don't have permission don't upload it. 

• Please note that the Library's print and electronic resources are not licensed for direct upload into GenAI tools. To stay compliant, avoid uploading journal articles or book chapters. Instead, we recommend providing the article title, DOI, or a direct link—this allows the GenAI tool to locate and summarise the content appropriately.

• The outputs you produce and intend to reuse could also be infringing copyright, if they are reproducing substantial parts of originals. Be mindful of this and read relevant indemnity clauses that could protect you to some extent.

• Read the terms and conditions of the tool carefully. Does the provider have a licence to reuse your input and generated output? If necessary, do not agree to your prompts being reused by the tool.

• Terms and conditions of the tool should also specify who owns the output. Please also be mindful of the fact that GenAI outputs are not necessarily protected in the first place.

• Consider the benefits of GenAI tools being trained on research produced by you. Read any relevant clauses or addenda related to reuse by GenAI in your publishing agreements and ask your editor questions if you have reservations.