Copyright Guide

About the License:
Mount St. Joseph University subscribes to the Annual Copyright License (ACL) from Copyright Clearance Center. This license provides Mount faculty, staff, and students the permission they need in order to share information from publications that are covered under the terms of the license.  The license increases what you already have access to through the Library homepage, thereby creating greater access to materials with copyright permissions.

What the License Covers:

• Comprehensive, institution-wide license allows employees of MSJ to use and share text-based copyrighted content, reducing need for pay-per-use permissions.
• Electronic course content (Blackboard, posting on password-protected myMount pages. internal email, distance learning)
• Coursepacks (print and electronic)
• Traditional and electronic reserves
• Research collaboration (faculty-to-faculty sharing)
• Scan paper copies of works when an electronic version is not available
• Extends to students studying abroad, satellite campuses, and distance education programs

What the License Does NOT Cover:

• Music
• Videos
• External promotion and advertising use
• Reproduction of the entire work unless specifically noted by rights holder
• Creation of a library, collection, or repository to substantially replace the institution’s need for a particular work or subscription

Verify Coverage:
Search RightFind Academic database to verify if a specific publication and type of use are covered under the license. See also the PDF Guide to Searching RightFind Academic. Other helpful resources can be found at the Copyright Clearance Center's "Tools and Resources -- Getting Started" page.

Cite Sources:

When you use a source that is covered by the Annual Copyright License, include the following citation on a cover page to indicate that you have confirmed copyright permission to post or distribute the content.

EXAMPLE:

“[TITLE] [Volume/Edition] by [Author]. Copyright [Copyright Year] by [Rightsholder].  Reprinted by permission of [Rightsholder] via Copyright Clearance Center.”

Resource Links:

Guidance Pertaining to Copyright Compliance and Generative AI Software

Generative AI (e.g., ChatGPT, CoPilot, Gemini, etc.) tools are impacting the academic and societal landscape. As a relatively new and evolving technology, regulations and legal guardrails are still being formed. At the same time, publishers are investigating instances of copyright violation and are filing lawsuits. The MSJ Library recognizes that you may not be fully aware of the legal ramifications of using copyrighted materials and licensed materials within a generative AI system. Therefore, this guidance serves to inform you of your responsibility to comply with US Copyright Law and with publisher/vendor license agreements that govern the Mount Community’s use of materials that are subject to these agreements:

• It is not permissible for you to upload any copyrighted or licensed materials (books, journal articles, etc.) to any publicly accessible or public-facing generative AI software unless you have express permission from the owner and/or licensor of these materials. Put differently, if uploading copyrighted/licensed materials to a generative AI system would result in uploading those materials to the corpus used to train that AI software, then you are not permitted to upload those materials to that system.
• Some licensors do allow users to upload their licensed materials to a local-use only instance (that is, an instance only available to the subscriber or authorized users—see further explanation below) of generative AI software that is in a self-hosted environment, provided the data processed by such software is not shared with unauthorized third parties. Before uploading any materials to a local-use instance, check to see if there is a license agreement for the materials, and check the terms of the agreement to make sure such use of the materials is permitted.

Please note that there are strong potential consequences for violations of publishers’ license agreements. Publishers can and will revoke access to their materials without warning for the entire Mount Community if they suspect a breach of their license agreements. If you elect to use generative AI, you are personally responsible for ensuring that you follow all publishers’ policies about use of their licensed materials with generative AI systems.

For more in-depth discussion of this topic, see further explanation below. Thanks in advance for being a responsible member of the Mount Community and a good steward of intellectual property by complying with US Copyright Law, as well as publishers’ license agreements regarding use of their content with generative AI.

If you have questions about the use of copyrighted or licensed content with ChatGPT or generative AI apps, please feel free to contact the MSJ Library: library@msj.edu.

­­­­­­­­­­­­­­­­­­­­_________________________________________________________

As stated above, it is not permissible for you to upload any copyrighted or licensed materials (books, journal articles, etc.) without the copyright owner’s and/or licensor’s express permission to any publicly accessible or public-facing AI software. So for example, you must not upload copyrighted or licensed content to the publicly accessible version of ChatGPT. This is due to the fact that the copyrighted/licensed material is being added to the generative AI’s training corpus, and thus you cannot control how the generative AI software uses such material. For example, ChatGPT could include (and in fact has been known to include) large excerpts of copyrighted material in the output that it creates, and to do so without attribution. This is why, for example, the New York Times (and other copyright owners) have filed lawsuits against ChatGPT.

It might be permissible for you to upload copyrighted material to your own local-use (private) instance of AI software for your own research use or the use of other authorized users. I say “might” here because there are caveats in this scenario you will need to consider. Many copyrighted materials are published with licensing agreements that contains terms of use. These terms of use specify both which uses of the licensed material are permitted and which uses are prohibited. A few publishers are now providing some limited allowances for uploading their content to a local-use instance of AI software for data analysis and other research purposes, as long as the publisher’s content is not being added to the corpus used to train that software. However, some publishers prohibit such use of their materials, and still other publishers have not yet specified in their license agreements whether such use is permissible or not. Therefore, if you considering uploading licensed materials to your own private instance of AI software, you will need to determine on a case-by-case basis whether or not you have express permission to do so based on the publisher’s license agreement, because you can only upload content from publishers that expressly grant permission to do so in their licensing agreements. Furthermore, before sharing any output created by your instance of the AI software with others, you will need to make sure the output doesn’t contain any unattributed excerpts from the licensed materials you loaded into the software. Finally, you may only share the output with authorized users as determined by the publishers’ licensing agreements. In this context, “authorized users” would be limited to only other members of the Mount Community (i.e., current students, employees, and authorized Mount-affiliated users). To be clear: the burdens of compliance, both with reference to copyright law and any relevant licensing agreements, is upon you as the end user.

To further clarify what a “local-use only instance means,” this would be an instance of a generative AI software program that is only available to the end user(s) and therefore the output isn’t publicly accessible. In other words, a local-use instance would be a purchased subscription to the software. So for example, ChatGPT has a subscription plan where you can purchase a subscription to ChatGPT for your workplace. In this example, only you and your fellow Mount Community members would have access to this instance of the software and therefore only you have access to its output. OpenAI, the owner of ChatGPT, wouldn’t be able to add the content added to your instance of the software to its training corpus, and therefore no one else in the world would have access to the output other than you and Mount colleagues and/or students (unless, of course, you chose to share it with someone else).

This is in contrast to the ChatGPT website. When you interact with the free version of ChatGPT on the website, anything you upload to ChatGPT on the public/free website goes into ChatGPT’s corpus, and the software trains on that content and can incorporate the content you uploaded into any output (documents, slides, etc.) that the software creates. Your content could theoretically be included in any output that ChatGPT creates for any other end user in the world. And thus this creates a copyright violation. For example, if I were to upload a PDF of a journal article to the free/public version of ChatGPT and tell it to give me a summary of the article, the full text of that journal article is now in ChatGPT’s training corpus, and ChatGPT can include any or all of the journal article in any other output it creates for any other end user in the world. Thus, I’ve shared a copy of that article with the rest of the world in a way that is a direct violation of US Copyright Law. In contrast, if I purchase a subscription to ChatGPT, I could possibly upload an article to my subscription/local-use instance and ask it to provide a summary of the article for me and whoever else is an “authorized user” of the subscription (like my fellow Mount co-workers or students, for example). In this scenario, the content of that journal article is, so to speak, “staying with” my own instance of ChatGPT—it isn’t being added to ChatGPT’s training corpus. Now notice I said I could “possibly” upload the article because I still need, even in this situation, to be aware that the publisher of that journal article may not allow its content to be uploaded to any instance of a generative AI software, either publicly or privately (i.e. subscriber-only) accessible.

In summary one of the key factors in determining whether or not you can upload copyrighted or licensed content to a generative AI software is this question: will the content you’re uploading be added to the corpus that trains whichever generative AI software you’re using? If the answer to that question is yes, then you cannot upload the content in question unless you first secure express written permission from the copyright owner and/or licensor to upload their content. If the answer to that question is no, you might be able to upload the content, but again, the responsibility is upon you as the end user to determine whether or not any licensing statements are in effect that would restrict uploading the material in question to any generative AI software, regardless of whether or not the material will be added to the AI’s training corpus. Again, this is due to the fact that some publishers provide permission for such activity, while others strictly prohibit any use of their materials with any instance of generative AI software, even local-use (private) instances.