Kyle McCarter has served in the United States Army for over 20 years. He presently works as a chief of staff in the Indo-Pacific area of responsibility. Divergent Options’ content does not contain information of an official nature, nor does the content represent the official position of any government, any organization, or any group.

Title: Assessment of Human-Technological Teaming Solutions on Intelligence Collection Limitations Across Africa’s Etymological Terrain

Date Originally Written: November 16, 2025

Date Originally Published: February 1, 2026

Author / Article Point of View: The author is a U.S. Army officer currently serving in the U.S. Indo-Pacific Command (INDOPACOM) area of responsibility. He previously served in the U.S. Africa Command (AFRICOM) area of responsibility where he controlled small unit Human and Counterintelligence Operations spanning the continent of Africa. The author believes linguist limitations across the U.S. Intelligence Community (IC) hinder operations in the AFRICOM region.

Summary: Despite changes in focus and resourcing by U.S. administrations on global threats, Africa remains a vital region for American national security interests. Timely and accurate intelligence collection remains a critical requirement to maintain national security. This article assesses how the U.S. IC might overcome linguistic barriers on the African continent.

Text: At 1.5 billion people, the continent of Africa ranks second most populous in the world and hosts the highest linguistic diversity with over 2,000 living languages [1]. The linguistically diverse African population stands astride approximately 30% of the world’s known mineral reserves, 8% of oil and 7% of natural gas reserves, and the largest cobalt, diamond, platinum, and uranium reserves in the world [2]. Africa plays a vital role in international relations due to its vast wealth of natural resources that remained untapped. Therefore, world powers such as China, India, Russia, Saudi Arabia, the European Union, and the United States all attempt to maintain influence in Africa.

Notably, the United States links the development and stability of Africa to its own national security—including efforts to stop the spread of extremist groups such as Boko Haram, Al Shabab, and ISIS. Timely and accurate intelligence collection, processing, and dissemination by the U.S. IC remains vital to ensuring positive development and stability in Africa. Although English is one of the three primary languages (English, French, and Arabic) spoken in Africa, intelligence collection is hampered by the vast number of local languages and dialects. The U.S. IC faces the task of overcoming language barriers to increase the volume and quality of information gathered and processed.

Lowenthal and Clark categorize the five main intelligence disciplines as Open Source (OSINT), Human (HUMINT) and Counterintelligence (CI), Signals (SIGINT), Geospatial (GEOINT), and Measurement and Signatures (MASINT) [3]. Of the five main categories, OSINT, HUMINT and CI, and SIGINT require an understanding of and the ability to use the local target language. GEOINT “consists of imagery, imagery intelligence, and geospatial information” and does not rely on an understanding of the local language [4]. Finally, “MASINT embodies a set of sub-disciplines that operate across the electromagnetic, acoustic and seismic spectrums, and material sciences ” and requires limited language abilities for collection [5]. While MASINT and GEOINT do not require language skills, they only answer a portion of the intelligence requirements and should not be relied on in full for Africa-related threat reporting and analysis. For the U.S. to gain an enriched understanding of the intelligence picture in Africa, the U.S. IC should utilize HUMINT and CI, SIGINT, and OSINT, all which require astute language abilities.

The time required to train a collector in a specific language is a major limiting factor for the U.S. IC – especially for languages in Africa. The Defense Language Institute Foreign Language Center (DFLIFLC) is the Department of Defense’s primary language training resource. The mission of the DLIFLC remains to “provide exquisite, regional and culturally based foreign language education globally in support of the Department of War [Defense] to provide a competitive edge for the joint force…” [6]. For Africa, the DLIFLC offers Modern Standard Arabic, Egyptian Arabic, and Sudanese Arabic but these courses are each 64 weeks long. Standard French is also offered at DFIFLC and requires 36 weeks to complete [7]. After a translator or intelligence collector has been trained in one of the main Arabic languages or standard French, further training is required for that individual to be able to communicate effectively with one of the 2,000 dialects on the African continent.

Because the U.S. IC has a limited number of collectors with the required language skill sets and extensive training timelines are often involved, the U.S. IC relies heavily on local national interpreters. The use of local national interpreters presents its own set of challenges to intelligence collection. Without proper vetting and screening by trained CI agents the interpreter could use the intelligence collector’s reporting to settle tribal, ethnic, or business disputes [8]. The use of local interpreters also limits the questioning ability of the collector due to classification concerns. Local national interpreters will have their own unique sets of loyalties, priorities, and limitations of their abilities.

Technology offers a significant force multiplier in the absence of a large cadre of trained linguists. As the African continent’s digital footprint expands, so do opportunities for information gathering. The advent of Neural Machine Translation (NMT) and Large Language Models (LLMs) is revolutionizing intelligence collection [9]. For example, the Intelligence Advanced Research Projects Activity (IARPA) continues development – since 2018- of a universal translator through its MATERIAL program to enable cross-language information retrieval and summarization for intelligence analysts [10] [11]. Simultaneously, grassroots efforts are flourishing across Africa. Initiatives like Izwe AI in South Africa, Sunbird AI in Uganda, and Lelapa AI’s InkubaLM are creating translation and transcription services for dozens of local languages, often using a “human-in-the-loop” model where local language specialists verify AI-generated translations to ensure accuracy and cultural nuance [12] [13] [14].

Although significant developments have highlighted the value of these tools, language remains an inherently human endeavor and cannot be fully replaced by technology [15]. The optimal solution, therefore, is not a purely technological one, but instead a “human-AI teaming” approach that creates a state of “superagency” for the intelligence collector [16]. To enable this teaming approach, the U.S. IC would train a small cadre of linguists with skills in targeted languages. They would act as the human element in a Computer-Assisted Translation (CAT) workflow, using technology to automate initial translations while providing the critical cultural and contextual understanding that machines currently lack [17]. Moreover, this approach would ensure that the final intelligence product is not only linguistically accurate but also culturally relevant, a crucial factor for effective operations in Africa’s complex information environment. In addition to conducting highly sensitive collection, this initial cadre could also inform the development and fielding of future technologies.

Overcoming the linguistic challenges facing the U.S. IC in Africa will not be accomplished with one single solution. Yet, the potential of human-AI teaming capabilities to transform intelligence collection in linguistically diverse locales should be thoroughly considered. With an already commanding position in global AI technology development, the U.S. stands poised to address its national security interests with greater effect than ever before.

Endnotes:

[1] Lewis, Paul M., Gary F. Simons, and Charles D. Fennig (eds.) (2016). Ethnologue: Languages of the World, Nineteenth edition. Dallas, Texas: SIL International. Retrieved November 16, 2025, from http://www.ethnologue.com.

[2] African Development Bank. African Natural Resource Center (ANRC). Retrieved November 16, 2025, from  https://safe.menlosecurity.com/doc/docview/viewer/docN3269321FBFD7a1bdfff05c889880b0f2cf5828fe1aa4811e161b2ec2bd36a82150ba79e0a240.

[3] Lowenthal, Mark, and Robert M. Clark (2016). The Five Disciplines on Intelligence Collection. Washing, DC: SAGE Publications, Inc.

[4] United States Code, Title 10, Section 467 (10 U.S.C. §467).

[5] Morris, John (1996). MASINT, American Intelligence Journal 17, no. 1 & 2: 24.

[6] Defense Language Institute Foreign Language Center. Retrieved November 16, 2025, from http://www.dliflc.edu/.

[7] Ibid.

[8] United States (2007). The U.S. Army/Marine Corps Counterinsurgency Field Manual. Chicago: University of Chicago Press: 3-134.

[9] Machine Translation: How It Works and Tools to Choose From. Retrieved January 12, 2026, from https://builtin.com/artificial-intelligence/machine-translation.

[10] Intelligence Community Working on a Universal Translator. Retrieved January 12, 2026, from https://www.meritalk.com/articles/intelligence-community-working-on-a-universal-translator/.

[11] MATERIAL: Machine Translation for English Retrieval of Information in any Language. Retrieved January 12, 2026, from https://www.iarpa.gov/research-programs/material.

[12] Izwe: Transform Audio and Video into Accurate Text Fast in all South African Languages. Retrieved January 12, 2026, from https://deepgram.com/ai-apps/izwe.

[13] Translation and Speech: African Language Technology. Retrieved January 12, 2026, from https://sunbird.ai/portfolio/african-languages/.

[14] Lelapa AI Launches Africa’s First AI Large Language Model. Retrieved January 12, 2026, from https://africaworld.princeton.edu/news/2024/lelapa-ai-launches-africa%E2%80%99s-first-ai-large-language-model.

[15] Human Translation Services. Retrieved January 12, 2026, from https://thetranslationcompany.com/translators/human-translators.htm.

[16] Superagency in the Workplace: Empowering People to Unlock AI’s Full Potential. Retrieved January 12, 2026, from https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work.

[17] Machine Translation vs. Computer-Assisted Translation. Retrieved January 12, 2026, from https://poeditor.com/blog/machine-translation-vs-computer-assisted-translation/.