AI Generated Content - University Marketing and Communications

UT Guidelines for AI-Generated Content

Overview

The University of Texas at Austin is committed to leading the responsible and innovative use of artificial intelligence (AI) generated content for marketing and communications teams across campus.

Our approach is guided by the University’s dual principles of being AI Forward and AI Responsible — empowering our community to explore the potential of AI while upholding the highest standards of ethics, privacy and academic integrity.

Updates

As AI technologies continue to evolve, we will regularly update the following guidelines to support responsible and effective use of AI in marketing and communications work. If you have any questions or feedback, please contact internalcommunications@austin.utexas.edu.

Guiding Principles

All individuals are 100% responsible for any output they create when using artificial intelligence.
All AI use must follow the Acceptable Use of AI Tools guidance and UT’s Data Classification Standard.

How Generative AI Can Support Communications Work

Appropriate Use Cases

Appropriate use cases require critical discernment of output. For example, it is acceptable to make a chart summarizing data, but you are responsible for making sure it is correct and accountable if it is not.

Examples of Appropriate Generative AI Use

Alt-Text & Plain Language: Improve readability and access.
Analytics: Summarize performance, data visualization (charts, graphs, etc.)
Brainstorming: Headlines, campaign ideas, story angles.
Design/Multimedia Support: Concept generation.
Drafting Content: Outlines, first drafts, timelines/calendars, memos, presentations, announcements, captions.
Editing Content: AI can assist with recommendations to help bring content into the brand voice (tone, clarity, AP style suggestions).
Editing Images: Cosmetic touch-ups, background simplification, cropping, resizing, color and lighting adjustments.
Web & UX: FAQs, chatbot responses, accessibility reviews.

All AI-generated content must be reviewed by a human before publication.

Do’s and Don’ts: Communications & Content

Do

Use AI for drafting, brainstorming and editing.
Use UT-approved tools to ensure data security.
Maintain a prompt library for consistency.
Use AI to enhance accessibility and efficiency.
Use AI to help you summarize and synthesize content for internal use.
Use AI to improve workflow and productivity.
Comply with all UT policies and applicable laws.
Vet all work for consistency, accuracy and alignment with the brand.
Engage in constant upskilling and development to ensure literacy and fluency with AI.
Practice responsible adoption principles endorsed by the University.
Protect the UT brand against AI slop — low quality content with no oversight and no genuine value to audiences.

Don’t

Don’t publish final content without careful human review.
Don’t input confidential or proprietary information (e.g., student records, internal strategy) when using free or non-UT-managed AI tools without proper authorization. UT-approved tools can be used for confidential or proprietary information.
Don’t fabricate quotes, testimonials or personal stories.
Don’t replace subject matter expertise with AI.
Don’t create misleading visuals or alter real subject matter (people, events, settings, etc.).
Don’t use AI to generate or modify UT trademarks or dilute the UT brand.
Don’t violate copyright or licensing rules.
Don’t create a representation (digital twin) of UT leadership or spokespersons.
Don’t use generative AI for proprietary research, legal advice, hiring decisions, grading or academic work unless explicitly permitted.
Don’t use AI for translation without full review by a native speaker for accuracy, tone and clarity.

Do’s and Don’ts: Graphics & Images

AI-generated visuals may support early-stage creative exploration. For example, use AI to generate sample images for wireframes, early-stage drafts or mockups. But have your final assets be brand-approved images, images from a photographer or purchased artwork.

If AI-generated visuals are published, they must meet institutional standards and include final review and approval by a human.

Do

Use AI for cosmetic touch-ups (e.g., stray hairs, shine reduction, object/blemish removal).
Use AI to prototype possible images, infographics or other media to use as a basis for your own creativity.
Use AI for background simplification (e.g., remove/adjust shadows or unwanted objects such as trash and construction debris).
Use AI for cropping, resizing, color and lighting adjustment.
Indicate when a published image is AI-generated rather than photographed or hand-created.
- Example: “Image created with [AI tool name]”

Don’t

Don’t use AI-generated images to manipulate, mislead or misinform.
Don’t generate images that appear photorealistic yet depict events or scenes that did not occur. This includes generating images of students, faculty or staff who do not “exist” on our campus.
Don’t create, modify or imitate University wordmarks, logos or any trademarked property.
Don’t depict future or speculative campus architecture, structures, art, monuments or grounds as if they currently exist.
Don’t modify existing campus structures to include non-existent features (e.g., adding floors or statues, changing signage or facades).
Don’t create AI slop — low quality images, videos or other content with no oversight and no genuine value to audiences.

Helpful Links and Resources

Understanding AI Limitations

While AI offers powerful capabilities, it also comes with risks and limitations:

Privacy & Security – Protect sensitive data; use only approved platforms.
Hallucinations – AI may generate inaccurate or fabricated content.
Misalignment – Outputs may not reflect UT’s values or goals.
Bias – AI can reflect or amplify societal biases.
Ethics – Responsible use is essential to maintain trust and credibility.
Cognitive Offloading – AI should support, not replace, critical thinking.

Best Practices

Additional Resources