Artificial Intelligence and Human-out-of-the-Loop: Is It Time for Autonomous Military Systems?
DOI: 10.46941/2024.2.18
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper examines the feasibility and implications of deploying fully autonomous, AI-driven military systems, specifically addressing the transition from "human-in-the-loop" to "human-out-of-the-loop" (HOOTL) operations. The research is motivated by demographic shifts, declining military enlistment, and the need to reduce civilian and soldier casualties. It analyzes the current state of unmanned military systems (UMS), the technical capabilities of artificial intelligence (AI), and the vulnerabilities these systems face from adversarial electronic warfare (EW) and AI-specific attacks. The study reviews the evolution of unmanned aerial vehicles (UAVs), noting the shift from large, expensive High-Altitude Long Endurance (HALE) systems to smaller, tactical UAVs, as demonstrated in the Russo-Ukrainian conflict. It also covers unmanned ground, surface, and underwater vehicles. The paper details AI technologies, particularly Convolutional Neural Networks (CNNs), which enable pattern recognition and decision-making. It identifies critical limitations in AI training, including data bias, overfitting, data scarcity, and lack of interpretability. Furthermore, it explores adversarial techniques used to disrupt AI systems, such as GPS spoofing, signal jamming, and the injection of noise or false data into training sets to create "adversarial examples" that cause misclassification. The findings indicate that while fully autonomous systems offer significant advantages in speed, scalability, and cost-efficiency, they remain vulnerable to sophisticated EW measures and AI-specific attacks. The paper argues that current AI models lack the reliability required for HOOTL deployment, particularly in distinguishing civilians from combatants, which poses severe ethical and legal risks. Consequently, the author concludes that "manned-unmanned teaming" (MUM-T) is the most viable near-term model. In this framework, humans retain central oversight ("human-on-the-loop"), delegating routine tasks to AI while maintaining control over critical engagement decisions. The paper recommends that future development focus on improving AI robustness against adversarial attacks and establishing a comprehensive legal framework to govern autonomous systems, rather than pursuing immediate full autonomy.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | Crossref | — | — | 1 | 2026-08-09 |
| archive | success | canonical_url | — | — | 1 | 2026-08-09 |
| extract | success | pdftotext | — | — | 4 | 2026-08-10 |
| clean | success | clean | — | — | 2 | 2026-08-10 |
| chunk | success | chunk | — | — | 2 | 2026-08-10 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 2 | 2026-08-10 |
| promote | success | — | — | — | 1 | 2026-08-09 |
| summarize | success | llm | qwen3.6-27b-nvidia | summ-v5 | 2 | 2026-08-10 |
| tag | success | vector_similarity | — | — | 17 | 2026-08-11 |
| verify | success | — | — | — | 2 | 2026-08-10 |
Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.