As Agentic AI systems begin to operate at scale across the Department of the Air Force (DAF)—from cyber defense to mission planning to software modernization—the central question is shifting. It is no longer, “Do we have a human in the loop?” but rather, “Do we have the right human in the loop?”

This session examines how to identify, train, and empower key personnel—Guardians, Airmen, and civilians—who are uniquely qualified to supervise Agentic AI teams. These AI “mission partners” can reason, plan, and act across complex workflows; however, without the correct human oversight they introduce real risks to provenance, governance, and trust. We will explore how AI power amplifies both human strengths and weaknesses. Agentic AI can decompose complex problems, generate technical artifacts, and propose decisions across cyber, operations, and IT domains, but that same power can magnify poor guidance, hidden biases, and weak controls. An unprepared human in the loop can unintentionally approve flawed outputs or break traceability between inputs and decisions, while a well-prepared AI supervisor can turn these systems into disciplined, transparent force multipliers.

The session will argue that “key personnel” is now a competency, not just a billet. The right person in the loop is not merely a subject-matter expert or generic “AI operator,” but a new hybrid profile—someone who understands mission context, technical stacks, and AI behavior well enough to frame clear intents and constraints, challenge AI reasoning (not just conclusions), interpret confidence signals, detect hallucinations, demand evidence, and preserve provenance and governance.

Finally, we will differentiate training for AI fluency from basic AI literacy. Power users of Agentic AI require tailored, role-specific development—not generic “AI 101.” We will present a practical framework that includes role-based competency models (e.g., AI Supervisor, AI Product Owner, AI Risk Officer), skill maps spanning mission expertise, data literacy, prompt engineering, V&V, governance, and progressive training pathways from foundational literacy for the broader force to advanced curricula for designated AI key personnel. In an AI-enabled future, mission success will depend on having the right human prepared and empowered as the final authority over intelligent machines.

Key Learning Objectives

By the end of this session, participants will be able to:

• Explain why “the right human in the loop” is essential for Agentic AI systems and how AI power can amplify both good and bad human guidance.

• Identify key competency characteristics for AI-critical roles (e.g., AI Supervisors, AI Product Owners, AI Risk Officers), beyond traditional subject-matter expertise or generic “AI operator” skills.

• Describe how to maintain provenance, governance, and trust by tying AI-assisted decisions to source data, models, human approvals, and applicable policy/ethics/RMF constraints.

• Differentiate AI-fluency from basic AI literacy and outline training strategies that build verification, validation, and governance discipline, not just tool familiarity.

• Apply a practical framework for workforce development to recognize who should be “in the loop” for Agentic AI in their units and design progressive, custom-tailored training pathways for those key personnel.