Studying the Use of Generative Artificial Intelligence in Undergraduate Research at the U.S. Military Academy

John Scudder¹, Hayden Deverill², James Bluman³, Rob Harrison⁴, Patrick Kelly¹, Jonathan Paynter³, and Kevin Scott¹

¹ U.S. Army

² U.S. Army Cadet Command

³ Department of Mathematical Sciences, U.S. Military Academy

⁴ Cyber Research Center, U.S. Military Academy

Download the PDF

Abstract

The recent surge in generative artificial intelligence (AI), including large language models like ChatGPT, creates both opportunities and challenges in postsecondary education. To examine how rapidly evolving technology shapes undergraduate research at the U.S. Military Academy, faculty from five academic programs launched a quantitative, descriptive, longitudinal survey that tracks cadet AI use, perceptions, and educational outcomes. Year one results show cadets experiment with AI across many research tasks and use varies by discipline: STEM cadets often lean on AI for code editing, while humanities cadets rely on it for writing revisions. These early insights reveal promising gains in productivity and idea generation but also raise concerns about overreliance and uncertain boundaries for the proper use of AI. Over the next three years, we will follow cadet behavior and attitudes to guide faculty development and instructional strategies related to AI use. Our primary goal is to understand how generative AI affects undergraduate research, and our secondary goal is to give educators data-driven guidance for ethical, effective, and pedagogically sound AI integration in courses.

The release of ChatGPT in November 2022 thrust generative artificial intelligence (AI) into mainstream use. The emergence of widely available free and low-cost tools to boost productivity and human creativity seems poised to impact most areas of society and industry (Eapen et al., 2023). Within higher education, generative AI has already disrupted common practices, particularly in the domains of academic integrity, instructional methods, and research practices (Gruenhagen et al., 2024; Nazaretsky et al., 2025). As generative AI continues to evolve, its impact on the educational landscape remains multifaceted and uncertain. This article introduces a quantitative, descriptive, longitudinal study that seeks to understand the use of generative AI in undergraduate research and independent study courses, with a focus on its adoption across five distinct disciplines: mathematics, computer science, philosophy, English, and social sciences. These courses often demand high levels of creativity, critical thinking, and original scholarly output—precisely the domains where generative AI can either amplify or truncate student learning and performance.

At the U.S. Military Academy (USMA) in West Point, New York, where academic integrity and pedagogical rigor are foundational, initial guidance on generative AI was introduced in 2023. While it emphasized the importance of integrity and instructor-specific policies, the use of generative AI remained decentralized and varied across disciplines (Reeves, 2023). Given this context, faculty members across five academic programs initiated a collaborative, multiyear research project to track the adoption, utility, and implications of generative AI in student-driven research activities. This study aims to document and analyze the technological engagement and ethical considerations among cadets from the class of 2024 through the class of 2029.

Literature Review

The Rise and Scope of Generative AI

Generative AI comprises a class of technologies that leverage vast sets of training data to generate novel content in formats including text, images, audio, and video (Feuerriegel et al., 2023). AI technology has quickly diversified, producing numerous tools including large language models, code generators, design and meeting assistants, and multimedia generators (Cao et al., 2023). These systems allow users to streamline iterative creative processes, aiding in everything from lesson planning to budget forecasting to music creation. While the productivity benefits of AI are clear, scholars and technologists alike express concern over issues of trust, the dilution of human creativity, and possible erosion of ambition or ingenuity (Wu et al., 2025).

Real-world examples illustrate the potential disruptive power of AI tools. A Colorado artist used Midjourney AI to win a digital art competition (Kuta, 2022), and deepfake videos threaten the perceived integrity of journalism (Scott, 2024). The sheer scale of content creation is staggering. In under two years, users have created over 15 billion AI-generated images using DALL-E, Stable Diffusion, Adobe Firefly, and Midjourney—matching the entire output of photography’s first 150 years (Valyaeva, 2023). The scope and magnitude of impacts from generative AI seem to vary among disciplines, institutions, and subject matter, with leaders in the AI development community like OpenAI responding by contributing research and new technologies to address potentially harmful uses, including forgeries (Metz & Hsu, 2024).

Generative AI in Higher Education

Higher education has responded to the use of AI with a mix of curiosity, skepticism, and experimentation. A 2023 Chronicle of Higher Education survey of 404 academic leaders revealed a dual sentiment: while many recognized the potential for AI to reduce educational inequities, they also worried about its implications for degrading academic integrity and authentic learning (Anft, 2023). Responses across academia ranged from outright bans (Nolan, 2023) to attempts at strategic integration (Coffey, 2024). Moreover, a worldwide survey of 450 secondary and postsecondary institutions indicated that, as of May 2023, only 10% of schools had provided formal guidance on the use of AI in educational settings (O’Hagan, 2023). Guidance in academia is missing despite the rapid adoption of the technology by an estimated 100 million users by May 2023, surpassing the adoption rate of social media applications such as Snapchat and Instagram (UNESCO, 2024). These sources reflect a recognition that, despite widespread adoption of generative AI within most cultures, pedagogical adaptation is lagging and still evolving.

Professional organizations such as the Association for Computing Machinery (2023) and the American Psychological Association (2023) have provided discipline-specific guidelines for the use of AI. Both organizations permit the use of generative AI in scholarly writing, provided it is properly disclosed and attributed. The American Association of Colleges and Universities is also launching an Institute on AI, Pedagogy, and the Curriculum to promote ethical and equitable integration (Kelly, 2024). Critics, however, caution that AI-enhanced efficiency can short-circuit essential cognitive processes in student learning (Lang, 2023, 2024).

At the institutional level, universities have developed various approaches to dealing with AI. Harvard University, for example, issued decentralized guidance, allowing individual faculty and departments to set specific AI policies (Harvard University Information Technology, n.d.). Purdue University took a more centralized and comprehensive approach in early 2024, covering AI syllabus language, detection tools, and copyright concerns (Rickus, 2024). Other schools, such as the University of Michigan and University of California San Diego, have gone further by developing proprietary generative AI tools for increased privacy, accessibility, and equity (Coffey, 2024). The University of Michigan’s AI tools report daily usage exceeding 15,000 users (O’Connell, 2024), yet few institutions have published detailed policies specific to using AI in research at the graduate or undergraduate levels. West Point took early action by publishing institution-wide AI guidance in advance of the 2023–2024 academic year. This guidance encouraged transparency by both students and faculty, discouraged faculty reliance on AI detection tools due to high false-positive rates, and emphasized faculty autonomy in AI policymaking (Reeves, 2023).

Student Attitudes and Usage Trends

Recent studies shed light on student usage and perceptions of generative AI tools. A Tyton Partners survey of 1,600 students across 600 institutions found that nearly half used generative AI tools, with 12% identifying as daily users (Shaw et al., 2023). Most users relied on AI for relatively low-complexity tasks such as summarization or paraphrasing. Another study, published in the International Journal for Educational Integrity, also focused on student attitudes toward generative AI and surveyed 2,500 students at the University of Liverpool in spring 2023. The study found that the majority of students were supportive or somewhat supportive of using tools like Grammarly, but 70% were unsupportive or somewhat unsupportive of using tools like ChatGPT to write entire essays, and more confident writers were less likely to use or consider using generative AI for academic purposes (Johnston et al., 2024).

Notably, prior literature also indicates that younger populations tend to adopt new technologies more readily than older cohorts (Eisma et al., 2004; Lam & Lee, 2006; Rogers et al., 2017). Research suggests that undergraduate students may be able to incorporate generative AI more rapidly than instructors or research advisors (Coffey, 2023). As such, we have a unique opportunity to learn from the student population about how generative AI can be used to accelerate research progress at the undergraduate level.

Motivation for the Current Study

In this rapidly evolving environment, the need to understand how generative AI is affecting undergraduate research is urgent. Research and independent study courses demand high levels of critical thinking and sustained scholarly effort, often culminating in summative outputs such as theses, software, or technical reports. As such, these academic spaces present a prime opportunity to assess how students might meaningfully integrate generative AI tools to enhance creativity, productivity, or both.

Faculty from five academic departments at West Point (Mathematical Sciences, Law and Philosophy, English and World Languages, Electrical Engineering and Computer Science, and Social Sciences) initiated a longitudinal study to document trends in generative AI use among undergraduates from the classes of 2024 through 2029. As shown in Table 1, student exposure to generative AI tools varies depending on their graduation year, creating a natural experiment for observing adoption patterns over time.

This study aims to provide real-time insight into how generative AI is shaping undergraduate scholarship across a wide range of academic contexts. By focusing on a young, relatively tech-savvy population engaged in intellectually demanding work, the research will illuminate not just patterns of usage but also student attitudes, ethical considerations, and discipline-specific challenges in adapting to one of the most transformative technologies in recent history.

Research Questions

The primary goal of our research is to understand the use of generative AI by students across multiple academic disciplines within the context of undergraduate research. The study strives to learn from West Point’s student population so that instructors have an accurate understanding of the use of generative AI. The knowledge from the research will inform instructors on how to best advise, teach, and demonstrate the use of generative AI at the USMA. The research includes a multiyear survey intended to answer the following questions:

How do students use generative artificial intelligence in their undergraduate research?

How does student AI usage differ across subject areas?

How does student AI use change over time as individuals and in the aggregate?

Are there clear trends in how the use of generative AI evolves over the course of the multiyear study?

This article describes the research design initiated in January 2024 and provides a snapshot of early results from one semester of research.

Methodology

Research Design

This multiyear project employs a longitudinal, quantitative, and descriptive-statistics survey design. For the first six months reported, the research team fielded a retrospective Qualtrics instrument to undergraduate cadets, asking them about their use of generative AI in research-focused courses. Because the literature on student use of large language model tools is still in its infancy, the study’s initial objective is to describe, rather than predict or explain, patterns of adoption, perceived benefits, and deterrents. The survey consisted of demographic and background information, categorical descriptors of generative AI use, free-text responses, and a section of questions associated with generative AI adoption.

The three-item scale for generative AI adoption in coding, based on the first two-month survey responses, showed acceptable internal consistency with a Cronbach’s alpha of 0.69 after standardizing the data from the item response scales (n = 71). Face and content validity were established through expert review by the authors, all of whom serve as undergraduate research advisors and routinely integrate AI tools into student projects. Each item was examined for clarity, alignment with the study constructs (e.g., perceived usefulness, academic-integrity concerns), and disciplinary inclusiveness. To further strengthen validity, the first administration of the survey functioned as a pilot study. The research team reviewed survey responses for ambiguous items. This iterative approach will continue each semester, enabling the instrument to evolve in response to emerging AI policies and classroom practices.

Research Setting

We conducted this study at the USMA. The academy enrolls approximately 4,400 undergraduate cadets annually and focuses on developing leaders of character who will commission as US Army officers. Cadets complete a broad core curriculum, choose from majors offered by 13 academic departments, and participate in collaborations with 27 research centers. The Cadet Honor Code upholds academic integrity, and the Documentation and Acknowledgment of Academic Work translates that code into concrete citation and collaboration guidelines. After USMA issued its initial guidance on generative AI in 2023, the dean’s office updated the Documentation and Acknowledgment of Academic Work to frame AI assistance as equivalent to help from a peer (Reeves, 2023).

The academy does not impose an institution-wide ban on generative AI; individual instructors decide whether cadets may use such tools. Cadets must acknowledge any AI assistance that materially shapes their work, yet they do not cite it in the same manner as a book or article (Reeves, 2023). Appendix A depicts a course-level policy memo that governs thesis work in data science, operations research, and mathematical sciences. Although we require every cadet to uphold the Honor Code, survey responses are anonymous. The consent statement clarifies that the study cannot link answers to individual students, and that participation carries no punitive consequences.

Initial Participants

The first survey of the study was launched in the spring of 2024. Between January and May, 236 cadets in 12 courses across five departments—English and World Languages (EN), Philosophy (PY), Computer Science (XE and CS), Social Sciences (SS), and Mathematics (MA)—were invited to respond. Table 2 shows the full distribution of invitations. All invitees took at least one undergraduate research course; most worked on a senior thesis or capstone, while the rest pursued independent study projects.

All participants provided informed consent before participating in the voluntary survey. The instrument asked students to enter their name and identification number, but the Office of Institutional Research removed all identifiers before we analyzed the data. The USMA Institutional Review Board reviewed and approved the study under protocol CA-2024-73.

Research Timeline

This longitudinal study is designed to run for at least three years. Each survey item is phrased to track how cadets’ attitudes toward generative AI shift over time and throughout their research journey. Because some second year cadets completed the first survey in 2024, their answers will be compared with those they provide as seniors. Free response prompts are also included to capture developments in the fast moving AI landscape. As academia adopts new generative AI tools, the survey will document their influence on undergraduate researchers.

Data Collection and Analysis

The study will field the survey instrument twice each semester—once during the first half and again after cadets submit final presentations or papers. Gathering data at two points allows researchers to track how generative-AI use evolves within a single course, especially as pressure mounts to finish final assignments. The study will keep the same cadence in future data collection efforts.

At each collection point, cadets receive a Qualtrics link. The instrument blends multiple-choice, free-response, and numerical survey questions. After cadets provide their consent, furnish their demographics, and select their major(s), branching logic guides them through prompts that match their earlier answers. Sample items include

• Does your research include any computer coding? Y/N
• What percentage of your code was generated using AI? Scale: 0-100%
• What percentage of your code was fixed/troubleshooted using AI? Scale: 0-100%
• If you used a coding copilot, describe how you employed it.
• When comparing the extent to which you trust the information, how would you compare Generative AI tools to other web sources (Stack Exchange, articles, papers, etc.)?
• Did you use generative AI in a way that you regret? If so, how?

After each survey, the Office of Institutional Research merges survey responses with institutional data, including course grades, and then removes all personal identifiers. The research team receives only anonymized records.

With each future survey, researchers will run descriptive analytics to identify trends on the primary axes of comparison:

• Changes in generative AI usage for an individual cadet over time
• Difference in usage between cadets from different graduating classes
• Variations between responses in different disciplines of study

Basic natural language processing techniques are applied to the free text answers. Bigrams and trigrams are extracted to spot recurring phrases that describe generative AI workflows, and sentiment analysis gauges overall attitudes toward generative AI.

The first year is used as an exploratory period and early findings will guide refinements to question wording, distribution methods, and branching logic. In years two and three, the study will compare both aggregate trends and each cadet’s longitudinal changes. The research will also add a short faculty module to monitor how instructors’ views on generative AI evolve and how they respond to student feedback.

Results and Discussion

Survey Respondent Demographics

Of those 236 students who received the survey during its first available semester, 71 completed it. Table 3 provides the class year of each survey participant; almost half (46.55%) of the participants are in the class of 2024, as most seniors are required to complete a research project as part of their coursework. Table 4 shows the academic major of each survey participant. Notably, 52.86% (37/70) of participants are STEM majors, and 47.14% (33/70) are non-STEM majors. Survey participants represent a diverse sample of all class years and types of majors, providing insight into a variety of generative AI use cases.

Selected Survey Results

This section includes the initial results after one semester of survey analysis. Table 5 presents the results specific to the use of generative AI in coding. Based on the 51 cadets who answered the question, an average of 25.10% of participants’ code was generated by generative AI, based on the mean of all of the question respondents. For both STEM and non-STEM majors, the initial survey results indicate that generative AI is widely used to assist with coding.

Table 6 shows the responses from cadets who used generative AI in their written products. For those cadets who answered the question, generative AI is used for idea generation, text generation, and proofreading. Our initial results suggest that generative AI could have significant applications in written research.

Table 7 shows the specific ways cadets used generative AI in their research. Of the cadets who answered the question, ChatGPT is the most prominent use for coding and written study.

Table 8 shows the responses regarding the perceived gains cadets achieved in their research by using generative AI. The responses suggest that generative AI improved more than half of the participants’ research productivity by 2.57, or 2.57 times more productive compared to when the student did not use generative AI.

Table 9 shows the responses to how using generative AI in research has helped cadets gain confidence in other professional contexts. Of those cadets, 67.27% (37/55) agreed or strongly agreed that using generative AI in research gave them confidence to use in different areas.

Discussion

The initial survey results from one semester show a clear student-perceived positive response to using generative AI in research efforts. These results may be swayed by a selection bias, where students who enjoy using generative AI are more likely to complete our survey. However, as an answer for research questions 1 and 2, the results indicate that regardless of the academic discipline, students experience a considerable boost in productivity when using generative AI to write and produce code. The first year of this quantitative, descriptive, longitudinal study reveals that cadets at the USMA are actively integrating generative AI into their research, with patterns of use varying across disciplines. While early findings suggest increased productivity and creativity when using AI, the study also highlights the potential issues of misuse or overdependence.

Conclusion

As we move into subsequent years of the study, researchers will continue to administer the same survey to collect information from students. The primary aim is to build a more comprehensive understanding of how generative AI influences undergraduate learning and research. The secondary aim of the study is to inform faculty development and contribute to a framework for the responsible, discipline-sensitive integration of AI tools in undergraduate research across higher education. The research will include our first analysis of longitudinal responses and enough data to dive more deeply into various statistical analyses.

---

References

 

American Psychological Association. (2023, November). APA journals policy on generative AI: Additional guidance. https://www.apa.org/pubs/journals/resources/publishing-tips/policy-generative-ai

Anft, M. (2023). Perspectives on Generative AI: College leaders assess the promise and the threat of a game changing tool. The Chronicle of Higher Education. https://connect.chronicle.com/CHE-CI-WC-2023-09-25-C-AI-CHE_LP.html

Association for Computing Machinery. (2023, April 20). ACM policy on authorship. https://www.acm.org/publications/policies/new-acm-policy-on-authorship

Cao, Y., Li, S., Liu, Y., Yan, Z., Dai, Y., Yu, P. S., & Sun, L. (2023). A comprehensive survey of AI-generated content (AIGC): A history of generative AI from GAN to ChatGPT. arXiv. http://arxiv.org/abs/2303.04226

Coffey, L. (2023, October 31). Students outrunning faculty in AI use. Inside Higher Ed. https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2023/10/31/most-students-outrunning-faculty-ai-use

Coffey, L. (2024, March 21). Universities build their own ChatGPT-like tools. Inside Higher Ed. https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2024/03/21/universities-build-their-own-chatgpt-ai

Eapen, T. T., Finkenstadt, D. J., Folk, J., & Venkataswamy, L. (2023, July 1). How generative AI can augment human creativity. Harvard Business Review. https://hbr.org/2023/07/how-generative-ai-can-augment-human-creativity

Eisma, R., Dickinson, A., Goodman, J., Syme, A., Tiwari, L., & Newell, A. F. (2004). Early user involvement in the development of information technology-related products for older people. Universal Access in the Information Society, 3(2), 131–140. https://doi.org/10.1007/s10209-004-0092-z

Feuerriegel, S., Hartmann, J., Janiesch, C. & Zscheck, P. (2023). Generative AI. Business and Information Systems Engineering, 66, 111–126. https://doi.org/10.1007/s12599-023-00834-7

Gruenhagen, J. H., Sinclair, P. M., Carroll, J. A., Baker, P. R. A., Wilson, A., & Demant, D. (2024). The rapid rise of generative AI and its implications for academic integrity: Students’ perceptions and use of chatbots for assistance with assessments. Computers & Education: Artificial Intelligence, 7, 100273. https://doi.org/10.1016/j.caeai.2024.100273

Harvard University Information Technology. (n.d.). Generative artificial intelligence (AI) guidelines. Retrieved 19 May 2024 from https://huit.harvard.edu/ai/guidelines

Johnston, H., Wells, R. F., Shanks, E. M., Boey, T., & Parsons, B. N. (2024). Student perspectives on the use of generative artificial intelligence technologies in higher education. International Journal for Educational Integrity, 20(1), 2. https://doi.org/10.1007/s40979-024-00149-4

Kelly, R. (2024, September 17). New AAC&U institute to explore challenges and opportunities of AI in teaching and learning. College Technology. https://campustechnology.com/articles/2024/09/17/new-aacu-institute-to-explore-challenges-and-opportunities-of-ai.aspx

Kuta, S. (2022, September 6). Art made with artificial intelligence wins at state fair. Smithsonian Magazine. https://www.smithsonianmag.com/smart-news/artificial-intelligence-art-wins-colorado-state-fair-180980703/

Lam, J. C. Y., & Lee, M. K. O. (2006). Digital inclusiveness—Longitudinal study of internet adoption by older adults. Journal of Management Information Systems, 22(4), 177–206. https://doi.org/10.2753/MIS0742-1222220407

Lang, J. M. (2023, October 16). John Dewey on artificial intelligence in education [Substack newsletter]. A General Education. https://jamesmlang.substack.com/p/john-dewey-on-artificial-intelligence

Lang, J. M. (2024, February 29). The case for slow-walking our use of generative AI. The Chronicle of Higher Education. https://www.chronicle.com/article/the-case-for-slow-walking-our-use-of-generative-ai

Metz, C., & Hsu, T. (2024, May 7). OpenAI releases “deepfake” detector to disinformation researchers. The New York Times. https://www.nytimes.com/2024/05/07/technology/openai-deepfake-detector.html

Nazaretsky, T., Mejia Domenzain, P., Swamy, V., Frej, J., & Käser, T. (2025). The critical role of trust in adopting AI powered educational technology for learning: An instrument for measuring student perceptions. Computers & Education: Artificial Intelligence, 8, 100368. https://doi.org/10.1016/j.caeai.2025.100368

Nolan, B. (2023, January 30). Here are the schools and colleges that have banned the use of ChatGPT over plagiarism and misinformation fears. Business Insider. https://www.businessinsider.com/chatgpt-schools-colleges-ban-plagiarism-misinformation-education-2023-1

O’Connell, A. J. (2024, February 7). How (and why) the University of Michigan built its own closed generative AI tools. EDUCAUSE Review. https://er.educause.edu/articles/2024/2/how-and-why-the-university-of-michigan-built-its-own-closed-generative-ai-tools

O’Hagan, C. (2023, May 26). AI: UNESCO mobilizes education ministers from around the world for a co-ordinated response to ChatGPT. UNESCO News. https://www.unesco.org/en/articles/ai-unesco-mobilizes-education-ministers-around-world-co-ordinated-response-chatgpt

Reeves, S. (2023, June). Documentation and acknowledgement of academic work. Office of the Dean, Academic Affairs and Registrar Services. https://s3.amazonaws.com/usma-media/inline-images/centers_research/west_point_writing_program/PDF/DAAW_2023.pdf

Rickus, J. (2024, January 8). Purdue issues spring 2024 guidance on AI use in teaching and learning; register for Jan. 11 TLCoP town hall discussion. Purdue University News. https://www.purdue.edu/newsroom/purduetoday/releases/2024/Q1/purdue-issues-spring-2024-guidance-on-ai-use-in-teaching-and-learning-register-for-jan-11-tlcop-town-hall-discussion.html

Rogers, W. A., Mitzner, T. L., Boot, W. R., Charness, N. H., Czaja, S. J., & Sharit, J. (2017). Understanding individual and age-related differences in technology adoption. Innovation in Aging, 1(suppl_1), 1026. https://doi.org/10.1093/geroni/igx004.3733

Scott, L. (2024, January 16). Deepfakes a “weapon against journalism,” analyst says. Voice of America. https://www.voanews.com/a/deepfakes-a-weapon-against-journalism-analyst-says-/7442897.html

Shaw, C., Yuan, L., Brennan, D., Martin, S., Janson, N., Fox, K., & Bryant, G. (2023). Generative AI in higher education: Fall 2023 update of Time for Class study. Tyton Partners. https://tytonpartners.com/app/uploads/2023/10/GenAI-IN-HIGHER-EDUCATION-FALL-2023-UPDATE-TIME-FOR-CLASS-STUDY.pdf

UNESCO. (2024, September 18). Less than 10% of schools and universities have formal guidance on AI. https://www.unesco.org/en/articles/unesco-survey-less-10-schools-and-universities-have-formal-guidance-ai

Valyaeva, A. (2023, August 15). People are creating an average of 34 million images per day. Statistics for 2024. Everypixel Journal. https://journal.everypixel.com/ai-image-statistics

Wu, S., Liu, Y., Ruan, M., Chen, S., & Xie, X. Y. (2025). Human-generative AI collaboration enhances task performance but undermines human’s intrinsic motivation. Scientific Reports, 15(1), 15105. https://doi.org/10.1038/s41598-025-98385-2

   

Appendix A

SUBJECT: Guidance on the use of Generative AI in MA498/MA499/MA491/MAx89

1. References.

Documentation and Acknowledgment of Academic Work (DAAW), June 2023, Office of the Dean, United States Military Academy.

Dean’s AY25 Generative AI Guidance

Association of Computing Machinery Policy on Authorship link

2. Purpose. The purpose of this memorandum is to provide guidance to cadets and faculty engaged in undergraduate research to ensure that we are aligned with Dean’s policies on the use of Generative AI. The use of AI tools including Generative AI is generally encouraged. Understanding how these tools can enhance productivity and research progress is important. However, cadets and faculty should seek to use AI tools without short-circuiting the process of learning—particularly the process of learning to write in a formal, technical style appropriate for upper level undergraduate or early graduate level work.

3. Policy. Cadets are welcome to utilize Generative AI, but assistance must be documented according to the DAAW. In other words, if you use AI tools in your work, you must cite it. Refined guidance is as follows:

Do Not:

Upload sensitive, proprietary, CUI, or classified data into an AI tool. Think carefully about uploading any data that you did not generate or that is not a publicly available data set.

Intentionally take any action that would discredit or harm the United States Army, USMA, or the Department of Mathematical Sciences.

Do:

Writing. Regarding assistance in writing your Preliminary Written Report, Interim Report, or Thesis: follow the Association of Computing Machinery guidelines

Coding. Regarding assistance received in coding: fully document and comment your code where you have received assistance from Generative AI in generating or troubleshooting code.

Other Assistance. Regarding assistance received in other research tasks such as brainstorming, experimental design, literature review, image generation or efficiently using LaTeX: acknowledge assistance IAW with DAAW (using an in-line citation and an Acknowledgment page).

Communicate openly and often with your advisor about how to use AI ethically and effectively as well as the implications and potential pitfalls of various AI uses.

Follow whatever Sharing and Publication Policy the authors of your AI tool have published, such as the OpenAI Policy.

Follow any data sharing agreements that you have in place.

If you wish to publish, be sure to understand the limitations and restrictions of your desired publishing venue.

NOTE: It will be difficult to comply with the dean’s policies on Generative AI without keeping good notes and records of where data, insight, code snippets, etc. came from. Be diligent in recording how progress on your research is made and be sure to generate your documentation while you are conducting research activities (brainstorming, coding, writing, data analysis, etc.).

Back to Top

June 2026