Using AI Safely In DoD And Federal Work

AI is becoming a routine part of Department of Defense (DoD) and federal work, including ranked alert queues, resupply forecasts, generated briefings, and vendor-supported AI services. These outputs can appear complete and authoritative before anyone has examined how the model produced the output, which data or source material it relies on, whether the environment is authorized, or who is accountable if the output is incorrect. This course teaches DoD and federal personnel, along with the contractors who support them, how to make defensible decisions about when, where, and how AI should be used. No prior AI or machine learning background is required. You'll learn to identify common AI use patterns, identify the people, systems, and operations that could be affected, and make an initial GO, CONDITIONAL-GO, NO-GO, or ESCALATE decision. From there, you'll assess the data and information the system uses, verify the authorized environment, apply the five DoD Responsible AI principles, define appropriate human review and decision authority, and manage risk using the National Institute of Standards and Technology AI Risk Management Framework (NIST AI RMF). The course also covers GenAI-specific risks, including hallucinations, prompt injection, data leakage, insecure outputs, and provenance and supply-chain concerns, along with the testing, monitoring, records-management, and acquisition practices needed to manage them. By the end, you'll produce a traceable AI Decision and Oversight Packet that documents the use case, controls, supporting evidence, accountable owners, and criteria for proceeding, stopping, or escalating. The packet is based on documented evidence rather than assumptions.