Artificial Intelligence and Human Interaction: How to Keep the Human in the Loop
Abstract
Army leaders are looking to procure and implement artificial intelligence (AI) technologies to solve a varietyof problems and enhance existing capabilities across multiple portfolios. While there are benefits to implementingnew technologies, including AI, there is often a major pitfall: the human factor as a user is consistentlyunderrepresented. This disparity between how AI-enabled systems are being acquired and how they should beacquired is often related to a gap in the development of systems not aligning with Human Systems Integration(HSI) best practices. The design of systems that facilitate human-agent learning requires further guidance. We usedata from the System for Award Management (SAM) along with discussions from subject-matter experts both ingovernment and industry to capture how AI-enabled systems are currently being procured by the Army. Thecombined results of the teams methodology revealed that there are varying understandings across the Army ofwhat an AI requirement is, and there are no obvious processes or specific AI acquisition guidelines that areuniversally followed when developing an AI requirement. It was also apparent that HSI was not always includedin requirements as required by Army regulations. This disparity appeared to have three major root causes:immaturity of DOD Army guidance, shortcomings in AI-related training for acquisition personnel, and anegligence surrounding the incorporation of HSI elements into Army requirements.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2022
- Accession Number
- AD1200526
Entities
People
- Ashley N. Gizas
- Benjamin R. Hill
- Dawn P. Patterson
- Megan Meisner
- Nicole Wilson
Organizations
- Naval Postgraduate School