AI Resources

AI Framework for Responsible Use 

The N.C. Department of Information Technology has developed the North Carolina State Government Responsible Use of Artificial Intelligence Framework, a living document for state agencies, designed to provide a comprehensive risk management approach to the use of AI.

Aligned with existing privacy laws and IT policies, the AI Framework applies to both existing and new uses of all AI. It consists of principles, practices and guidance to provide a measured and consistent approach for state agencies to support innovation while reducing privacy and data protection risks.

AI Assessment

The Privacy Threshold Analysis (PTA) is used to assess privacy and security. It addresses the use of AI and requires a description of the project or system that identifies the proposed use of AI and how that aligns with the state's principles of responsible use. The PTA is used to analyze risks for projects or systems using AI or generative AI from a privacy perspective. Privacy risk assessment is based on the Fair Information Practice Principles (FIPPS), which underpin NCDIT’s Principles for Responsible Use of AI, a foundational component of the North Carolina State Government Responsible Use of Artificial Intelligence Framework.

State agencies seeking additional information about the PTA or privacy AI risk assessment can send an email to the Office of Privacy and Data Protection. Those seeking support to mature generative AI use cases can also send an email to the AI Working Group.  

Use of Publicly Available Generative AI 

Ethical and responsible use of publicly available generative AI requires alignment with the state’s policies, mission and goals. These guidelines are human-centered with a focus on uses of AI that benefit North Carolinians and the public good. This requires assessment of publicly available generative AI tools and use cases to ensure that any tool used by the state is trustworthy AI. 

Identifying AI Risks and Facilitating Responsible Deployment of Gen AI

Identifying & Mitigating Bias in AI

NIST SP 1270, Towards a Standard for Identifying and Managing Bias in Artificial Intelligence: Identifies three categories of bias in AI — systemic, statistical and human — and describes how and where they contribute to harms; describes three broad challenges for mitigating bias — datasets, testing and evaluation and human factors — and introduces preliminary guidance for addressing them.

Other Resources

AI Ethics & Strategy

AI News

Foundational Models

Foundation model acceptable use/Terms of Service policies:

AI Training

The N.C. Department of Information Technology has compiled the following resources from around the internet for those seeking to understand more about AI.

CourseDescription
Introduction to Generative AIGoogle's course focused on explaining what generative AI is, its uses and the differences between generative AI and traditional machine learning methods
What Is Generative AI?Microsoft's course explaining the basics of generative AI – including what it is, how it works, creating your own content, models, predictions for the future of AI and ethics
Your Guide to Generative AILearn Prompting's free course on AI that contains more than 60 content modules on how to safely and effectively use AI
Introduction to Responsible AIGoogle's free eight-hour course that includes information about responsible AI and why it is essential
Data Science: Machine LearningHarvard's eight-week (free to audit) course covering the basics of machine learning, how to perform cross-validation to avoid overtraining, some popular machine learning algorithms, how to build a recommendation system, and more
Chatting as a Productivity ToolMicrosoft's course covering how its chat tool can perform tasks and help streamline your workflow; addresses idea generation, data summarization and more 
Microsoft Azure AI Fundamentals: Generative AI Covers understanding how large language models form the foundation of generative AI; describes how Azure OpenAI Service provides access to generative AI technology; explains how generative AI applications support efficiencies; describes how prompts and responses can be fine-tuned; and addresses responsible AI principles
Generative AI Learning Plan for Decision MakersAmazon's three-hour course designed to introduce generative AI to business and technical decision makers; includes an overview of generative AI and an approach to plan a generative AI project
Generative AI for EveryoneCoursera's six-hour course covering how generative AI works, what it is, common use cases and limitations
Generative AI with Large Language Models Coursera's 16-hour course that provides foundational knowledge, practical skills and a functional understanding of how generative AI works