Brand Standards: AI
Best Practices for Generative AI Use
Adherence to Brand Standards and Legal Compliance
- Follow USU brand standards and respect copyright, trademark and intellectual property when using GenAI tools.
- Do not use AI tools to create, recreate, upscale or alter any university logos, spirit marks, or trademarked works.
- Avoid generating images or design elements with the intention to imitate or copy authentic assets.
Authenticity and Transparency
- Use GenAI with careful consideration of authenticity, accuracy, and transparency regarding university's communications with the public and stakeholders.
- Prioritize creating and using authentic content. Avoid altering the original intent of images.
- Disclose the use of novel generated media, especially if it is photorealistic or mimics authentic audio/video recordings.
- Maintain transparency, protect private data and avoid bias when using GenAI.
Appropriate AI Applications
- Use AI for brainstorming, summarizing or simplifying complex materials for understanding or rough draft transcription. Ensure all AI-generated content is reviewed by a human before finalizing.
- Generating transcriptions from audio and video is acceptable, provided they are reviewed by a human.
- AI may assist in creating contextual elements into a social media post, such as titles, captions, graphics, or sound.
Review and Quality Assurance
- Review all statements, body copy, or expanded information by a human before finalizing.
- Always have professional visual quality assurance when working with designs, images, or video.
- Continuously review AI generated marketing plans, recommendations, and summaries for market research and analytics.
Ethical Use and Limitations
- Do not use AI to alter or manipulate information, audio, video, or other media with the intent to mislead. When in doubt, obtain consent from the original author.
- Do not fabricate complete outcomes for social media content.
Technical Considerations
- If developing web code, work in a language you know and understand. AI generated code should always be reviewed and tested by human experts.
- Pay particular attention to security and privacy/sensitive data issues.