Military Review

 

Publishing Disclaimer: In all of its publications and products, Military Review presents professional information. However, the views expressed therein are those of the authors and are not necessarily those of the Army University, the Department of the Army, or any other agency of the U.S. government.

 

Using Open Access AI to Create Military Training from POW Experiences

Col. John P. Albano, MD, U.S. Army, Retired

Cmdr. Steven E. Linnville, PhD, U.S. Navy, Retired*

Lt. Jacob R. Westerberg, PhD, U.S. Navy

Travis V. Meyer

 

Download the PDF Download the PDF

 
U.S. prisoners of war stand in the yard at Hanoi’s Nga Tu So prison during an inspection visit in March 1973 by international observers and invited Western journalists

U.S. prisoners of war stand in the yard at Hanoi’s Nga Tu So prison during an inspection visit in March 1973 by international observers and invited Western journalists. The prison, surrounded by a high brick wall topped with barbed wire, had been a government building before the Vietnam War. Each cell housed about a dozen prisoners of war, and beds consisted of wooden platforms with mats. (Photo by Horst Faas, Associated Press)

ChatGPT is the flagship product of the American artificial intelligence organization OpenAI. As a natural language processing program, ChatGPT leverages artificial general intelligence to generate complex informational responses based on user input.1 When prompted to describe artificial general intelligence (AGI), ChatGPT stated,

In simple terms, Artificial General Intelligence (AGI) refers to highly autonomous systems or machines that possess the ability to understand, learn, and apply knowledge across a broad range of tasks at a level comparable to human intelligence. Unlike specialized or narrow AI, which is designed for specific tasks, AGI aims to exhibit general cognitive abilities similar to those of humans. It involves machines that can adapt, reason, problem-solve, and learn in diverse situations, much like how humans can apply their intelligence to various activities without being limited to a specific domain. AGI represents a level of artificial intelligence where machines can handle complex tasks across different domains, demonstrating a versatile and human-like understanding of the world.2

Reminiscent to “HAL” 9000 (a.k.a. Heuristically Programmed Algorithmic Computer) in Stanley Kubrick’s 1968 movie 2001: A Space Odyssey, AGI has made the leap from science fiction to science fact.3 The goal of OpenAI is to provide society cutting-edge, interactive computer programs that are safe and beneficial to use by anyone.4 One way this technology may be useful in a military environment is in developing timely educational curriculum. The application of this technology to support education has already been examined in a military context by way of graduate medical education.5

This study explores the use of AGI for curriculum development to prepare military personnel to survive a captivity experience. The data used in this study were from books published by former U.S. prisoners of war (POWs) held in Vietnam between 1964 and 1973. These POWs represented all branches of U.S. military service, most of whom were aviators. Their experiences included prolonged torture, malnourishment, and periods of isolation.

Of these POWs, many wrote and published works about their captivity experiences and their life experiences afterward. The advancement of AGI has provided a novel avenue to compile their published experiences to identify teaching objectives that could be used by military members through a method of instruction on survival and resilience. Here, survival is defined as “the state or fact of continuing to live or exist especially in spite of difficult conditions,” and resilience is defined as “the process and outcome of successfully adapting to difficult or challenging life experiences, especially through mental, emotional, and behavioral flexibility and adjustment to external and internal demands.”6 For the purposes of this study, survival is considered an ongoing process during an event and resilience is a process that can occur both during and after an event. These are two critical themes necessary to train military personnel in preparation for combat and life post-combat.

Method

This analysis included works exclusively written by former POWs about their captivity experiences in Vietnam. For this pilot study, four published books were selected.7 Approximately 80 percent of each book was digitally scanned and uploaded into the freely available ChatGPT version 3.5. Due to text analysis limitations per a single command prompt, segments of each book were uploaded separately. For each segment, ChatGPT was asked to identify ten common themes from the text. Once a set of common themes were developed for each book, all forty themes from the four books were inputted again into ChatGPT. The program was then prompted to develop a set of twelve common themes across them. Next, from these twelve themes, an intermediary set of themes were developed—one set on survival and another set on resilience. From these latter two themes, teaching objectives were developed focused specifically on survival and resilience.

Results

Table 1 (at the end of the article) shows the themes generated from the writings (the “narrators” are the four authors). The themes gathered include military training, internal and external struggles, comradery with other POWs, cultural and language differences with their captors as barriers, the constant fear and survival they had, not knowing when and if they would be released, and reflections at the time of writing and while in captivity.

From these twelve themes, two sets of themes were prompted, one set for survival (table 2) and one for resilience (table 3), with ChatGPT adding a thirteenth theme in resilience. Both the survival and resilience themes indexed both internal (memory, emotions, identity, and instinct) and external (physical, informational, and unity/support) themes (tables 2 and 3). These intermediary themes were generated for the final stage, which was to develop teaching objectives in these two areas (table 4).

In a side-by-side comparison of the survival and resilience objectives, the differences between them are in bold and italicized. For each of the thirteen themes, the objectives would require students to understand a difference between the two objectives and discuss in detail each of these objectives using these perspectives. The goal would be for students to develop a deeper level of understanding of these areas and how to reflect on and address them in their own learning process.

Discussion

This study was the first of its kind to use AGI technology to derive a set of common themes for captivity-related curriculum development. The major contributions of this study follow:

  • First, it is the only research to date to leverage AGI to understand the experience of captivity among former American POWs from their perspective.
  • Second, this work demonstrated the ability for AGI to generate domain-specific curriculum based on user inputs relevant to a military context. The use of this novel technology could be leveraged for content generation and instructional design to support instruction to military personnel as part of their military training. This has implications for military schoolhouses responsible for teaching survival, evasion, resistance, and escape curriculum.
  • Third, from a methodological perspective, this study has outlined a process to analyze voluminous works quickly with face validity from a cohort to produce shared themes across individuals with a shared experience. Future research in this area should consider cataloging all available works from cohort members for inclusion into data analysis. This additional effort will likely generate more comprehensive and nuanced insights. This study showed both the relevance and promise of leveraging novel AGI technologies to support military-related curriculum development. Moreover, such use of AGI is a major technological advancement in the twenty-first century and the way ahead for civilization.

Notes

  1. Dinesh Kalla et al., “Study and Analysis of Chat GPT and Its Impact on Different Fields of Study,” International Journal of Innovative Science and Research Technology 8, no. 3 (March 2023): 827–33, https://liapers.ssrn.com/sol3/liapers.cfm?abstract_id=4402499.
  2. “ChatGPT,” OpenAI, accessed 7 March 2024 [login required], https://chat.openai.com/.
  3. Aaron M. Lamb, “Through the Lens of HAL 9000: Using Stanley Kubrick’s 2001: A Space Odyssey as a Modeling Tool to Create a Precursive Sapient Quotient to Foster Humanity’s Moral Obligation to Evolve into Machines” (PhD diss., Johns Hopkins University, 2022), https://jscholarship.library.jhu.edu/items/9e87418f-d4a6-4bd9-88b0-8f0e9e35e687.
  4. “OpenAI Charter,” OpenAI, accessed 7 March 2024, https://openai.com/charter.
  5. Jonathan R. Spirnak and Sameer Antani, “The Need for Artificial Intelligence Curriculum in Military Medical Education,” Military Medicine (website), 20 October 2023, https://academic.oup.com/milmed/advance-article/doi/10.1093/milmed/usad412/7326531.
  6. Britannica Dictionary Online, s.v. “survival,” accessed 13 March 2024, https://www.britannica.com/dictionary/resilience; “Resilience,” American Psychological Association, accessed 7 March 2024, https://www.apa.org/topics/resilience.
  7. Everett Alvarez Jr. and A. S. Pitch, Chained Eagle: The Heroic Story of the First American Shot Down Over North Vietnam (Lincoln, NE: Potomac Books, 2005); Charles J. Plumb, I’m No Hero, a POW Story As Told to Glen DeWerff (self-pub., 1995); John Borling, Taps on the Walls: Poems from the Hanoi Hilton (Chicago: Master Wings Publishing, 2013); Frank Anton and Tommy Denton, Why Didn’t You Get Me Out? A POW’s Nightmare in Vietnam (New York: St. Martin’s Paperback, 2000).
 

Table 1. ChatGPT Facilitated Common Themes Analysis
(Table by author)

AI-Table1

Table 2. ChatGPT Facilitated Common Themes on Survival Analysis
(Table by author)

AI-Table2

Table 3. ChatGPT Facilitated Common Themes on Resilience Analysis
(Table by author)

AI-Table3

Table 4. ChatGPT Facilitated Common Teaching Objectives on Survival and Resilience Analysis
(Table by author)

AI-Table4

Table 4. ChatGPT Facilitated Common Teaching Objectives on Survival and Resilience Analysis (continued)
(Table by author)

AI-Table5
 

Col. John P. Albano, MD, U.S. Army, retired, is the program director for the Robert E. Mitchell Center (REMC) for POW Studies at Naval Air Station, Pensacola, Florida. He holds an MD from the University of South Dakota, an MPH from the University of Texas, and is board certified in aerospace medicine. He ended his twenty-five-year Army career as the command surgeon for the U.S. Army Space and Missile Defense Command. He has served at the REMC since 2012.

 

Cmdr. Steven E. Linnville, PhD, U.S. Navy, retired, received his psychology PhD from Southern Illinois University, Carbondale. His twenty-six-year Navy Medical Service Corps career involved HIV research on military performance, studying prolonged low-frequency sonar effects, teaching at the Army’s West Point as a Navy assistant professor, and exploring psychological resilience among Vietnam-era U.S. POWs at the Robert E. Mitchell Center, Pensacola. Post-retirement, he consults on cognitive performance and aviator health. *He is the primary author of this article.

 

Lt. Jacob R. Westerberg, PhD, U.S. Navy, serves as associate director and research psychologist at the Robert E. Mitchell Center for POW Studies at Naval Air Station, Pensacola, Florida, where he leads research on resilience, team dynamics, and leadership among former U.S. prisoners of war to support warfighter development.

 

Travis Meyer, MBA, has worked for the Department of Defense for the past eighteen years in both civilian and military roles. He is currently office manager at the Robert E. Mitchell Center for POW Studies at Naval Air Station, Pensacola, Florida, and a key contributor in the development of this article.

 

Back to Top

May-June 2024