Please note that on this page, Artificial Intelligence (AI) primarily refers to the new set of AI tools built upon large language models (LLMs).

Please see Artificial Intelligence – IT Purchase Process | Vendor Risk & License Management for a list of approved AI tools. 

NC State embraces the responsible and effective usage of artificial intelligence (AI) to elevate our missions and goals. Importantly, we believe that AI should support, but not replace, human involvement in decision-making and outputs. These guiding principles serve as a flexible framework for appropriate use regarding AI adoption, usage, and development. These principles will be reviewed and updated as needed. 

1. Ensure AI Tools are Approved Before Use: Several AI Tools are now available for use with green (not sensitive), yellow (moderately sensitive) and red (highly sensitive) data campus-wide. Use of University Data with unapproved AI tools requires an IT Purchase Compliance (ITPC) request. 
2. Authentication: NC State credentials must be used for all approved AI tools; personal accounts are not allowed to be used with university data. Using your NC State account ensures that your data is not used to train public AI models. Exceptions must be submitted for review as part of the ITPC process.
3. Privacy/Security: When using AI tools, be sure to follow NC State’s Data Management Framework and the IT regulations and rules, including but not limited to the Endpoint Protection Standard (RUL 08.00.18). AI tools operating with university data, regardless of deployment or execution method, should be secured using a combination of administrative and technical safeguards consistent with the applicable NC State data classification and associated privacy or security requirements. 
4. Transparency: Promote transparency around the data, methods, logic, and algorithms used in training and development of AI tools and models to help users and stakeholders understand how AI decisions are determined.
5. Attribution: NC State expects that attributions of AI use be applied in compliance with any applicable standards, including but not limited to requirements from funders, journals, syllabi, and other contracts.  
6. Quality and Accuracy: Modern generative AI tools can create content that is not truthful (aka hallucinations). Ensure any content you create using AI tools is high-quality and accurate. Verify the accuracy of AI outputs using additional sources or checks. 
7. AI Bias: When using AI tools, recognize that AI outputs are potentially erroneous and biased based on the context in which it is trained, its source data, its developers and its original intended purpose. When possible, strive to use AI tools built using data practices and algorithms that mitigate bias. When developing your own AI, ensure that quality controls such as “bias checks” or “code review” occur.
8. Accountability/Integrity: Use and develop AI tools in a manner that would be regarded as ethical, honest, and professional, meeting the compliance obligations and PRRs of NC State University. 
9. Skill Development and Human Training: Ensure that all people interacting with the selected AI, from its development through its implementation, are aware of both the possibilities and limitations of AI use. 

Note: For AI guidance specific to research, please see the following: Research Subgroup Statement on Responsible AI Use in Research

These guiding principles are provided and managed by the NC State University Artificial Intelligence Advisory Group.