Generative AI (GenAI) presents a number of challenges related to intellectual property and copyright. Please take the following considerations into account when using copyrighted material:
- Adding information you’ve written or created to a GenAI tool: When you upload or enter information in a GenAI tool, you may be giving the tool's developer a broad license to use, reproduce, and even train their AI models with your content, potentially leading to a loss of confidentiality. Read the terms and conditions carefully.
- Adding or uploading information written by someone else to a GenAI tool: You might be violating copyright or licensing agreements if you upload copyrighted material (such as a PDF of a journal article) to a GenAI tool. When Brandeis Library subscribes to a database, journal, or other online resource on behalf of faculty, staff, and students, we agree to the resource's licensing agreements, which protect the intellectual property of the authors represented in the resource. Database vendors often prohibit sharing their resources with those who do not have authorized access, which includes most GenAI tools, as users generally cannot control how an AI tool will use input once uploaded.
- Training data for Large Language Models: GenAI tools may have been trained on copyrighted material and other intellectual property without obtaining the consent of authors and other creators.
- Copyright of AI output: This is an ongoing discussion that will be decided in courtrooms in the US and other countries around the world. Copyright has typically been granted only for content created by humans.
- Infringement risk: GenAI models are trained on vast datasets, and sometimes they can derive closely from copyrighted works within their training data. This means an AI-generated output, even if it looks new to you, could inadvertently infringe on an existing copyrighted work, and you may be held liable for copyright infringement.
