Fundamentals of Generative AI

AI Is More Than ChatGPT. Government Needs More Than Prompts.

Artificial intelligence is changing how public agencies communicate, analyze information, manage programs, and serve the public. Yet using AI effectively requires more than learning how to operate a chatbot.

It requires judgment.

Fundamentals of Generative AI Compliance provides public-sector leaders, managers, project teams, and employees with a practical foundation for understanding, evaluating, and governing generative AI.

Written in clear, accessible language, this second edition explains how generative AI works, where it may be useful, and what government agencies must consider before placing it into practice. Readers learn to distinguish promising technology from unsupported claims—and responsible innovation from avoidable risk.

Inside, you will learn how to:

• Understand generative AI, machine learning, neural networks, transformers, and large language models
• Write more purposeful prompts and critically evaluate AI-generated information
• Identify risks involving accuracy, bias, privacy, security, accessibility, and misinformation
• Distinguish enacted law from proposed legislation, executive direction, agency policy, official guidance, and voluntary standards
• Evaluate AI across its complete lifecycle—from proposed use and procurement through testing, deployment, monitoring, correction, and retirement
• Establish meaningful human oversight, public disclosure, documentation, and accountability
• Use practical governance questions, worksheets, and templates to support agency decision-making

The second edition also includes a structured Learning and Application Pathway and Final Applied Project. Together, they help readers assess what they have learned, apply the material to a realistic public-sector use, and prepare for further study or certification without presenting the book itself as a professional license or government-recognized credential.

You do not need to become an artificial intelligence expert. You do need enough knowledge to ask better questions, recognize risk, evaluate evidence, and know when specialized review is required.

AI systems will continue to change. The responsibility to govern them will remain human.