AI Usage Guidelines

Updated: October 31, 2024

Generative AI, at its core, refers to a class of artificial intelligence systems capable of producing original and innovative outputs, such as text, images, audio, code, and more, in response to given prompts. These systems are trained on vast datasets and learn the patterns, styles, and structures present in the data, enabling them to generate content that closely resembles human creations. One prominent example of generative AI is the Language Model (LLM), such as OpenAI's GPT-4. LLMs are sophisticated models designed to generate human-like text by predicting word sequences. They operate by processing and analyzing large amounts of text data, learning grammar, context, and semantics. LLMs can be fine-tuned for specific tasks, making them versatile tools for tasks like content creation, translation, summarization, and even conversation.

Considerations while Using Generative AI

Ethical Considerations
Verify responses
Generative AI uses information from many sources, including sources with unreliable or unverifiable information. Also, LLMs, like ChatGPT, are trained to predict the next word in a sentence based on the text that has already been written. Responses from AI systems can produce what is currently termed “hallucinations,” or answers that are truly generated without corresponding source material–that is the content may be erroneous. Always validate information generated by AI services to assess that it is factual before using.

Understand that Generative AI is biased
Because AI systems can refer to historical information without context, some answers from AI systems may be biased. Caution should be used when using AI systems for decisions with ethical implications.

Recognize and respect copyrights
Generative AI systems, including those that generate text and images, leverage the work of others. This can sometimes create copyright concerns, especially if a content creator has removed their source material from consideration after creating a derived work. Include labels or context with anything published and derived from generative AI.

Security and Privacy Considerations
Protect Private Information
Use caution with private information. Information considered sensitive or confidential by University Policy 808 (https://policy.rice.edu/808), including private information about individuals and protected research data, should not be used with consumer-focused or publicly available AI services, including any AI plugins or addons included with other publicly available services. Information shared with these types of services may be exposed to unauthorized parties.

Review the Security of Services Before Use
Before licensing new services, including Generative AI services, contact the Information Security Office. The Information Security Office will perform a thorough review of the service to ensure that the service has security in place to maintain the protection of sensitive and confidential information appropriately.

Additional Resources
AI systems will impact teaching and learning in many ways, presenting both opportunities and challenges for the ways our course instructors and students engage in learning.

• AI and Teaching
• Example Syllabi Statements from CTE
• Fondren AI LibGuide

Please note that these guidelines are not new University policy; rather, they leverage existing University policies. As Generative AI is a rapidly evolving technology, we are committed to staying up-to-date on the latest advancements and updating our feedback accordingly.

Service*

*Service providers continue to evolve their contract Terms of Service and Compliance attestations. This chart is revised regularly to include updated information from each listed service provider. See previous versions. If you have specific questions about use cases, please contact OIT using the Request Help form.

Items in bold in the table below indicate that they were updated since the last version.