1. Home ›
  2. Policies and Guidelines ›
  3. Guidelines for Campus Communicators on Generative AI Use

Guidelines for Campus Communicators on Generative AI Use

Generative artificial intelligence tools—e.g., Google Gemini, ChatGPT, and Adobe Firefly, among many others—have the potential to help campus communicators do their work more effectively, quickly, and thoroughly. These tools also have several ethical questions inherent to their use. These guidelines, provided for all Notre Dame employees who communicate on behalf of the University, are meant to help staff navigate the ethical, legal, and practical uses of generative AI (GenAI) tools.

The full scope of campus AI use and approved tools can be found on the AI@ND website at ai.nd.edu.

What is generative AI?

AI has been around in many forms for years. The term includes everything from spell check to Google searches to Photoshop tools.

Generative AI is a term specific to the relatively new tools that create: With a prompt, users can have these tools write an essay, put together photo captions and alt text, develop a story outline or interview questions, draft a social media campaign, or create images. These tools can also be used to modify photos and videos, create podcasts from whole cloth, pull together executive summaries of lengthy reports, and complete many other tasks that would otherwise fall to an individual.

GenAI can be a great time-saving tool and brainstorming partner, and can help jump-start a project or bring cohesion to one. Think of this technology as a tool to help your work, not as a replacement for it. GenAI can help you work more efficiently, quickly, and innovatively—but the tools must be used responsibly.

AI@ND has published guidelines for using generative AI at Notre Dame. These guidelines are the basis for the guidance below. Please use the link to review the AI@ND guidelines if you plan to use GenAI tools.

AI@ND Guidelines 

How the AI@ND guidelines apply to campus communications

The guidelines below are intended to help communicators navigate their use of generative AI in a thoughtful manner that is appropriate for the University. When using AI tools in your work, please reference this page to ensure that you keep these considerations in mind.

Do:

  • Feel free to use generative AI in your work as desired, as outlined in these guidelines.

  • Disclose to your audience when generative AI was used in a substantial way—e.g., when an image was created with ChatGPT, a text was translated to another language, or AI was used to create a video voiceover. AI use does not need to be disclosed for things like story outline generation or transcription. Some sample language is below.

    • If an image was created with GenAI: This graphic/illustration/image was created by [name of staff member] using [name of generative AI program].

    • If an existing photograph or image was substantially altered with AI: This photograph from [name, outlet/company] was altered by [name of staff member] using [name of generative AI program]. [If relevant to keep credibility, briefly list what was changed.]

    • If text was translated using generative AI: This text was translated from [language] to [language] using [name of generative AI program]. Please report any errors to [email].

  • Be cautious when creating with AI, particularly with images—the use of AI can implicate third-party interests, such as copyright, rights in personality, or other intellectual property rights, that could expose the University or individual users to legal risk.

  • Be aware that all generative AI tools have inherent biases based on the human biases from the data they’re built on. When using these tools, watch for potential problems with language and representation.

    • Because many large language models have been built on essentially the entire internet, any existing human biases could come out in generative AI responses—as one example, a service could create an image of a Black man when it’s asked to create an image of a criminal.

    • For more information, see the AI@ND website, this MIT resource, or this Bloomberg article.

Do not:

  • Use AI-generated text verbatim. This applies to most contexts, with some exceptions.

    • AI-generated text is acceptable in specific uses such as alt text or non-user-facing web applications, though it still must be reviewed by a person.

    • AI-generated text is not acceptable if it’s used in a way that makes an audience believe it came directly from a University communicator.

  • Use AI to write or rewrite entire stories.

  • Use any AI-generated content that was not first reviewed and fact-checked by a human.

  • Include internal, sensitive, or highly sensitive data in AI prompts or uploads, unless the OIT has specifically approved the tool for it.

  • Use wholly AI-generated content (such as an image or video) without a disclosure.

  • Use AI to substantially alter any part of a photo or video, even with a disclosure.

    • Customary editing of a photo or video—for example, color correction, retouching, or other enhancements—does not need to be disclosed.

    • Do not use AI tools to change an image to the point of misrepresentation of what actually happened, or to alter what someone says in a video interview, as examples.

Remember:

  • Do not put confidential or non-public University information into generative AI tools, unless they have been specifically approved by OIT. There is no guarantee of privacy or confidentiality.

  • Text generated by AI can be inaccurate or misleading. You still must do your own fact-checking and research.

  • Even if you’re not generating content yourself, be aware that anything you find online (a stock image, a graphic, etc.) could potentially be AI-generated. Do your due diligence to avoid resharing false information.

  • Consider whether using generative AI is the best option for outward-facing content. It has many advantages, but unnecessary use of GenAI risks having a negative impact among certain audiences.

    • Do you need a photo or image? First check photos.nd.edu or images available on stock photo services or in the Creative Commons.

    • Do you need a logo or graphic? First check in with Notre Dame Creative.

    • Studies have shown that audiences often react negatively to AI-generated content when they know it’s AI-generated. Know how your audiences perceive AI-generated content and whether its use is beneficial.

Our credibility is rooted in our integrity. Any use of AI that undermines the University’s trustworthiness cannot be permitted.

Resources