





































Education, Language and Sociology Research 

ISSN 2690-3644 (Print) ISSN 2690-3652 (Online) 

Vol. 6, No. 4, 2025 

www.scholink.org/ojs/index.php/elsr 

80 
 

Original Paper 

A Comparative Study of Human and AI-Generated Spanish 

Translations of the Huangdi Neijing: Implications for Cross-

Cultural Transmission in the Age of Artificial Intelligence 

Yao Yan¹ & Yuanyuan Zuo1* 

¹ School of International Education, Yunnan University of Chinese Medicine, Kunming, China 

* Corresponding Author: Yanyan Zuo, School of International Education, Yunnan University of Chinese 

Medicine, Kunming, 650500, China. E-mail: 410506504@qq.com 

 

Received: November 18, 2025   Accepted: December 3, 2025  Online Published: December 12, 2025 

doi:10.22158/elsr.v6n4p80      URL: http://dx.doi.org/10.22158/elsr.v6n4p80 

 

Abstract 

The Huangdi Neijing is one of the foundational classics of Traditional Chinese Medicine (TCM) and 

plays a central role in the global dissemination of Chinese medical culture. With the rapid advancement 

of AIGC (AI-Generated Content) technologies, machine-assisted translation has become increasingly 

relevant to the translation of culturally and conceptually complex medical classics. This study conducts 

a comparative analysis of a representative Spanish translation of the Huangdi Neijing and its AI-

generated counterpart, using Reiss’s translation criticism framework. Under an IMRaD structure, the 

paper examines semantic accuracy, terminology choices, grammatical features, and cultural 

expressiveness to evaluate the strengths and limitations of human and AI translations. 

The results show that while AI translations demonstrate high efficiency and structural consistency, they 

often lack precision in medical terminology, contextual interpretation, and cultural connotations. Human 

translators, in contrast, excel at accurately conveying TCM concepts, interpreting classical Chinese 

syntax, and compensating for cultural gaps, although at the cost of longer production time and 

occasional inconsistencies. The study argues that human–AI collaboration offers an effective model for 

translating classical TCM texts, combining efficiency with cultural and conceptual fidelity. It concludes 

that human creativity and professional expertise remain indispensable in the translation of medical 

classics, especially in ensuring terminological accuracy and preserving humanistic value in cross-

cultural communication. 

 

 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

81 
Published by SCHOLINK INC. 

Keywords 

Huangdi Neijing, Traditional Chinese Medicine, Spanish translation, AIGC, machine translation, 

human–AI collaboration, translation criticism 

 

1. Introduction 

Traditional Chinese Medicine (TCM) has experienced renewed global attention over the past several 

decades, driven by increased international exchange, greater recognition of complementary and 

integrative medicine, and the expanding availability of translated classical texts. Among these, the 

Huangdi Neijing (Yellow Emperor’s Inner Canon) represents one of the core foundations of the TCM 

intellectual tradition. As both a medical and philosophical classic, its conceptual system—including Qi, 

Yin–Yang, the Five Phases, and the correspondence between humans and nature—presents unique 

challenges for translation into languages that lack equivalent epistemological frameworks. The Spanish-

speaking world, where TCM has enjoyed rapid growth since the late twentieth century, now has an 

increasing demand for reliable and readable translations of TCM classics. This makes Huangdi Neijing 

translation studies not only an academic inquiry but also a practical necessity for medical education, 

clinical training, and cultural exchange. 

The translation of classical TCM texts has historically relied on human translators with strong 

interdisciplinary backgrounds in sinology, medicine, and linguistics. However, recent developments in 

AIGC (AI-Generated Content) technologies have introduced new possibilities for accelerating translation 

processes and enhancing accessibility. AI translation tools—particularly large language models such as 

ChatGPT—are increasingly used to handle complex linguistic tasks. Their capability to process vast 

textual inputs, generate coherent outputs, and standardize terminology raises the question of whether 

machine-generated translations can meaningfully support or complement human translation of culturally 

and conceptually dense texts such as the Huangdi Neijing. Despite these emerging opportunities, 

empirical studies evaluating the performance of AI systems in translating TCM classics remain limited. 

The challenges of translating the Huangdi Neijing are well documented. The text is written in Classical 

Chinese, characterized by high contextual density, polysymy, implicit logic, and layered metaphorical 

structures. TCM terminology itself is culturally embedded rather than strictly biomedical; many key 

terms do not have direct equivalents in Western medical discourse. Terms such as “发陈 (fa chen),” “少

长  (shao chang),” and “精气  (jingqi)” require not only linguistic decoding but also conceptual 

interpretation rooted in the traditional Chinese medical worldview. Previous research (e.g., Tang, 2015; 

Yan, 2023; Hu & Wang, 2024) has emphasized the literary nature, cultural load, and semantic 

indeterminacy of TCM terminology—dimensions that often fall beyond the current capability of AI 

translation systems. Consequently, a rigorous comparison between human translators and AI-generated 

outputs becomes essential for understanding the strengths, limitations, and potential roles of AI in this 

specialized field. 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

82 
Published by SCHOLINK INC. 

In addition, the Spanish-speaking world presents its own translational landscape. Existing Spanish 

versions of the Huangdi Neijing—whether direct translations from Chinese or mediated through English 

and French—display considerable variation in terminology, annotation practices, and interpretive 

frameworks. The uneven quality of translations, the coexistence of literal and adaptive approaches, and 

the market presence of annotated or modernized interpretations further complicate the reception of the 

classic. With the rapid diffusion of AI translation tools among students, practitioners, and educators of 

TCM in Spain and Latin America, an evaluation of AI performance becomes both timely and necessary. 

Given this context, the present study adopts Reiss’s translation criticism theory as its analytical 

framework to systematically compare human and AI translations of selected passages from the Huangdi 

Neijing. Reiss’s model, emphasizing both linguistic and extralinguistic factors, provides a comprehensive 

lens through which semantic accuracy, terminological fidelity, stylistic features, and contextual 

appropriateness can be examined. Using a representative Spanish translation published by JG Ediciones 

(2005) as the primary human-translated source and ChatGPT-generated Spanish renderings as the AI 

counterpart, this research assesses the extent to which AI can approximate, support, or diverge from 

human interpretive processes. 

This study contributes to the field in three ways. First, it provides empirical evidence concerning the 

strengths and weaknesses of AI translation in a highly specialized cultural-medical domain. Second, it 

sheds light on the cognitive and interpretive gaps between human translators and AI systems when 

dealing with Classical Chinese and TCM conceptual structures. Third, it argues for a human–AI 

collaborative model in the translation of classical texts, advocating for an approach that leverages AI 

efficiency while preserving the human translator’s cultural sensitivity, professional judgment, and 

interpretive depth. 

By situating the discussion within the broader context of global TCM communication, this research 

underscores the continuing importance of translation quality in shaping international understanding of 

Chinese medical thought. As TCM gains visibility worldwide, accurate and culturally informed 

translations of classics such as the Huangdi Neijing remain indispensable for fostering intercultural 

dialogue, ensuring professional training standards, and facilitating the global dissemination of traditional 

medical knowledge. 

 

2. Literature Review 

Research on the translation of Traditional Chinese Medicine (TCM) classics has expanded substantially 

in recent decades, driven by growing global interest in TCM theory, clinical practice, and cultural 

heritage. Early scholarship focused on philological challenges, particularly the ambiguity of Classical 

Chinese grammar and the absence of direct equivalents for key TCM concepts in Western languages. 

Studies by Unschuld, Sivin, and others emphasized the epistemological gap between ancient Chinese 

medical cosmology and modern biomedical frameworks, arguing that translation must preserve 

conceptual integrity rather than impose Western interpretive categories. 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

83 
Published by SCHOLINK INC. 

In Spanish-language contexts, translations of the Huangdi Neijing remain limited. Existing versions vary 

in accuracy and philosophical depth, often relying on paraphrasing or explanatory notes to accommodate 

readers unfamiliar with TCM theory. Recent studies highlight the need for terminological consistency 

and methodological rigor in translating foundational TCM texts for international audiences. 

With the rise of artificial intelligence, research on AI-assisted translation has grown rapidly. Large 

language models such as GPT have demonstrated remarkable linguistic fluency but continue to struggle 

with domain-specific knowledge and culturally embedded concepts. Prior studies show that AI performs 

well in surface-level semantic tasks but lacks the conceptual grounding required for medical-

philosophical texts such as the Huangdi Neijing. As such, there is increasing academic interest in hybrid 

human–AI workflows that combine efficiency with interpretive depth.  

This study contributes to the literature by providing the first comparative analysis of human and AI-

generated Spanish translations of the Huangdi Neijing, highlighting their respective strengths and 

limitations and proposing a collaborative model for future translation practice. 

 

3. Methods 

3.1 Research Design 

This study adopts a qualitative comparative research design to evaluate the performance of human and 

AI-generated Spanish translations of selected passages from the Huangdi Neijing. Given the text’s 

complexity, Classical Chinese conciseness, and culturally embedded conceptual system, a qualitative 

analytical approach provides the most suitable means of examining semantic accuracy, terminological 

interpretation, stylistic coherence, and cultural fidelity. Reiss’s translation criticism framework is 

employed as the theoretical foundation due to its systematic categorization of linguistic and 

extralinguistic factors, allowing for a comprehensive and multilayered analysis of translation quality in 

both human and AI outputs. 

3.2 Materials and Data Sources 

3.2.1 Human Translation Sample 

The primary human-translated text used in this study is Su Wen: Canon de medicina interna del 

emperador amarillo (preguntas sencillas), published by JG Ediciones in 2005. This translation, widely 

circulated in Spain and Latin America, is produced by Julio García, a translator known for multiple TCM-

related works. It represents one of the most influential Spanish renderings of the Huangdi Neijing and 

thus provides a suitable benchmark for evaluating human interpretive practices, terminology choices, and 

annotation strategies. 

3.2.2 AI Translation Sample 

AI-generated translations were produced using ChatGPT (GPT-4.1 architecture, 2024 version), a state-

of-the-art large language model capable of multilingual text generation. For consistency and replicability, 

the same Classical Chinese source passages were input manually without additional context, prompts, or 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

84 
Published by SCHOLINK INC. 

explanatory notes. The output Spanish translations were collected and archived for systematic 

comparison with the human-translated versions. 

3.3 Selection of Textual Samples 

Given the length and thematic diversity of the Huangdi Neijing, the study focuses on representative 

passages from the Su Wen that pose distinct challenges for translation. These include: 

1) Passages involving TCM core concepts, such as Yin–Yang, seasonal correspondence, and jingqi 

(essence and Qi). 

2) Texts containing Classical Chinese syntactic ambiguity, multifunctional particles, or polysemous 

terms. 

3) Sections in which human translators provided additional annotations, explanations, or interpretive 

expansions—allowing comparison between intentional human guidance and AI’s autonomous output. 

This sampling strategy ensures coverage of key linguistic and conceptual difficulties inherent in the 

translation of TCM classics. 

3.4 Analytical Framework 

The study follows Reiss’s translation criticism model, which consists of two major dimensions: 

3.4.1 Intralinguistic (Linguistic) Factors 

⚫ Semantic accuracy: Precision in conveying original meaning, avoiding mistranslation. 

⚫ Terminological fidelity: Handling of culturally loaded TCM terms with no direct equivalents in 

Spanish. 

⚫ Syntactic structure: Correctness and readability of sentence formation. 

⚫ Stylistic appropriateness: Consistency with the tone and rhetorical features of Classical Chinese. 

3.4.2 Extralinguistic (Contextual) Factors 

⚫ Cultural interpretation: Ability to represent TCM worldview, metaphors, and analogies. 

⚫ Reader orientation: Whether translation choices accommodate target readers’ background 

knowledge. 

⚫ Purpose and function of the text: Faithfulness to the pedagogical or explanatory intent of the 

original classic. 

⚫ Translator’s role: Presence of annotations, interpretive notes, or culturally adaptive explanations. 

Through this dual framework, the analysis identifies systematic differences in the interpretive pathways 

of human and AI translations. 

3.5 Procedures of Analysis 

The analysis proceeded in four steps: 

1) Text alignment: Human and AI translations were aligned segment by segment for direct 

comparison. 

2) Unit-level analysis: Key translation units (terminology, metaphors, ambiguous structures) were 

identified. 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

85 
Published by SCHOLINK INC. 

3) Evaluation using Reiss’s criteria: Each unit was assessed for semantic accuracy, conceptual 

interpretation, stylistic adequacy, and cultural fidelity. 

4) Synthesis of findings: Patterns of strength and limitations in both translation modes were 

summarized into thematic results. 

This structured approach ensures transparency, replicability, and methodological rigor. 

3.6 Ethical Considerations 

No human participants were recruited for this study, and all materials used (published texts and AI 

outputs) fall under fair academic use. The study complies with ethical guidelines for research involving 

publicly accessible sources and technological tools. 

3.7 Limitations 

Two limitations should be acknowledged: 

1) Sample size: Only selected passages were analyzed, which may not represent every nuance of the 

Huangdi Neijing. 

2) AI model variability: Different AI models or versions may generate different translations; thus 

results reflect the behavior of the specific model used. 

Despite these limitations, the study provides a robust comparative evaluation that can inform future 

research on machine translation in culturally specialized domains. 

 

4. Results 

The comparative analysis revealed clear differences between the human and AI-generated Spanish 

translations of the Huangdi Neijing. These differences extended across semantic accuracy, terminology 

handling, syntactic and stylistic features, and cultural interpretation. The key findings are summarized 

below. 

4.1 Semantic Accuracy 

4.1.1 Human Translators Demonstrated Greater Sensitivity to Classical Chinese Syntax 

In several passages, the human translator showed an ability to interpret Classical Chinese particles and 

discourse markers that carry rhetorical rather than literal meaning. For example, in the sentence: 

“夫四时阴阳者，万物之根本也” 

the human translator misinterpreted 夫 as “man” (“para el hombre”), but ChatGPT correctly rendered it 

as a discourse particle: 

AI translation: 

“Los cuatro tiempos y el yin y yang son la raíz de todas las cosas.” 

This suggests that AI, relying on probabilistic linguistic patterns, sometimes outperforms human 

translators in identifying non-referential grammatical markers. 

4.1.2 AI struggled with contextual meanings and classical conceptual terminology 

Conversely, AI frequently misinterpreted lexemes requiring contextual medical knowledge. For instance: 

“夏为寒变，奉长者少” 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

86 
Published by SCHOLINK INC. 

AI interpreted “长” as “elderly individuals,” producing: 

“la salud de los mayores se verá afectada” 

while the human version correctly identified “长” as “growth”: 

“lo que se transformará en frío en verano; esto se llama ‘escasa entrega a crecer’.” 

This indicates that AI tends to default to the most statistically common meaning rather than infer 

contextual interpretations central to TCM theory. 

4.2 Terminology Handling 

4.2.1 Human Translators Preferred Domestication with Explanatory Annotations 

In cases where direct Spanish equivalents were unavailable, the human translator often provided: pinyin 

forms, descriptive glosses and conceptual explanations. 

This approach supports comprehension among readers lacking background in TCM. For example, human 

translators routinely added bracketed explanations after major TCM concepts. 

4.2.2 AI Favored Foreignization with Minimal Explanatory Support 

AI typically retained the pinyin term without interpreting its meaning or functional role within TCM 

theory: 

Example: 

“春三月，此谓发陈” → 

AI output: “En los tres meses de la primavera, llamados ‘fa chen’.” 

Without explanatory notes, the translation lacks interpretive depth and fails to convey the essence of the 

concept. 

4.2.3 AI Occasionally Produced Overly Literal Terminological Mappings 

For example, AI rendered “五藏盛” as “los cinco órganos están plenos,” which is semantically correct 

but lacks the dynamic sense of “optimal functional state” implied in TCM. 

Human translations, though irregular at times, conveyed richer interpretive nuance. 

4.3 Stylistic and Grammatical Features 

4.3.1 Human Translations Showed Stylistic Adaptation but Inconsistent Precision 

An example appears in the human rendering of age-related physiological decline, where entire segments 

were paraphrased to align with Western biomedical logic. This improved readability but introduced 

interpretive shifts: 

“sus riñones pierden su capacidad de hospedar líquidos…” 

The added explanation reflects a biomedical perspective rather than Classical Chinese conceptual logic. 

4.3.2 AI Translations Were Grammatically Stable but Stylistically Rigid 

AI generally produced structurally correct sentences but lacked stylistic variation. Its output tended to 

be: monotonic, literal and lacking rhetorical sensitivity.  

For a literary-philosophical classic such as the Huangdi Neijing, this rigidity diminishes expressiveness. 

 

 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

87 
Published by SCHOLINK INC. 

4.4 Cultural and Contextual Interpretation 

4.4.1 Human Translators Exhibited Greater Cultural Contextualization 

Human translators provided extensive cultural interpretation, including: referencing TCM classics (e.g., 

Nan Jing), elaborating relationships between organs and functions and maintaining thematic coherence 

with TCM pedagogy. 

Example: 

In explaining organ interactions, the human translator inserted clarifying commentary, which, though not 

in the original text, aids comprehension. 

4.4.2 AI Displayed Limitations in Cultural Insight 

AI mistranslated culturally dense or polysemous terms when biomedical analogues did not exist. For 

example: 

“人怀性情，故云应人。” 

AI’s translation: 

“corresponden a la naturaleza humana.” 

This rendering frames the phrase in emotional or psychological terms, whereas the original refers to 

individualized physiological responses. 

4.4.3 AI Was Susceptible to Character Recognition and Historical-Language Errors 

AI occasionally misinterpreted: rare characters, loan characters and variant forms.  

This is seen in unrelated—but representative—classical passages such as the misreading of “飧” and 

“疟”, showing that AI models still face challenges with archaic Chinese glyphs. 

4.5 Summary of Comparative Findings 

 

Dimension Human Translation AI Translation 

Semantic 

accuracy 

High in conceptual passages; 

occasional misreadings 

Correct for simple syntax; weak on 

contextual meanings 

Terminology Explanatory, reader-friendly Literal, minimal explanation 

Style Adaptive, sometimes interpretive Rigid, structurally uniform 

Culture High contextualization Limited interpretive depth 

Errors Pinyin mistakes; inconsistency Conceptual misinterpretation; 

character confusion 

 

 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

88 
Published by SCHOLINK INC. 

4.6 Overall Interpretation 

The results show that: 

⚫ AI excels at structural and basic semantic tasks 

⚫ Human translators excel at conceptual fidelity and cultural interpretation 

Neither approach alone meets the full demands of translating a culturally saturated medical-philosophical 

classic. Instead, the findings strongly suggest that: 

A human–AI collaborative model offers the most promising path for future translations of TCM classics. 

 

5. Discussion 

The comparative findings highlight both the potential and the limitations of AI-assisted translation in the 

highly specialized domain of Traditional Chinese Medicine (TCM). While AI translation tools have 

achieved substantial progress in handling general linguistic tasks, their performance with culturally 

embedded, conceptually dense classical texts remains markedly inferior to that of experienced human 

translators. Drawing on the results, this section discusses the implications of human–AI differences for 

TCM translation practice, cross-cultural knowledge transmission, and the evolving role of translators in 

the AIGC era. 

5.1 Divergent Interpretive Mechanisms Between Humans and AI 

The human translator and the AI model demonstrated fundamentally different interpretive mechanisms. 

Human translation relied on: 

⚫ Domain knowledge of TCM theory 

⚫ Familiarity with Classical Chinese textual logic 

⚫ Sensitivity to cultural metaphors 

⚫ Continuous inferential reasoning 

AI translation, by contrast, operated through: 

⚫ Statistical pattern prediction 

⚫ Lexical association 

⚫ General linguistic structure mapping 

These divergent mechanisms explain the observed discrepancies. AI handled surface-level semantics 

with ease but consistently struggled with deeper conceptual interpretation. This aligns with prior research 

suggesting that large language models often reproduce patterns without genuine semantic grounding, 

especially in domains where meaning is constructed through cultural schemas rather than explicit 

discourse structures. 

The Huangdi Neijing is not merely a medical manual; it encodes a cosmological view of the body, nature, 

time, and society. Such conceptual systems do not have direct equivalents in Western epistemology, 

making them inaccessible to AI when explicit contextualization is absent. 

 

 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

89 
Published by SCHOLINK INC. 

5.2 The Role of World Knowledge in Translating TCM Concepts 

One of the clearest patterns that emerged is that TCM translation requires more than bilingual proficiency; 

it requires integrated world knowledge. For example: 

⚫ The AI model misinterpreted “长” as “elderly individuals,” demonstrating a lack of medical-

philosophical background. 

⚫ Human translators, although occasionally inconsistent, were able to reconstruct meaning aligned 

with the TCM worldview. 

TCM terminology often functions within a network of conceptual relations, such as: 

⚫ The correspondence between organs and natural elements 

⚫ Dynamic processes like “opening–closing–pivoting” (开合枢) 

⚫ The cyclical movement of Qi 

AI lacks stable representations of such systems, which leads to semantic flattening and conceptual 

simplification. This finding reinforces the argument that translation of classical medical texts is 

inseparable from cultural and theoretical understanding. 

5.3 Cultural Fidelity vs. Linguistic Efficiency 

The contrast between human and AI translations reflects a broader tension between cultural fidelity and 

linguistic efficiency. 

⚫ Humans provided richer contextual notes, cultural interpretation, and conceptual nuance, but 

needed more time and introduced occasional surface-level errors. 

⚫ AI delivered syntactically precise and rapid output, but lacked interpretive depth and produced 

terminological mismatches. 

This divergence suggests that neither method alone is sufficient. AI’s efficiency is valuable, especially 

for initial drafts or large-scale batch translation. However, cultural fidelity—essential for accurate 

transmission of TCM knowledge—cannot be achieved by AI alone. 

This echoes concerns in translation studies that excessive reliance on machine translation risks 

marginalizing the translator’s interpretive agency and may result in the erosion of humanistic value in 

the translated text. 

5.4 Human–AI Collaboration as an Emerging Translation Paradigm 

The results support a hybrid translation model in which human translators and AI tools complement each 

other’s strengths. In such a model: 

⚫ AI generates preliminary drafts, handles repetitive content, or assists with lexical consistency. 

⚫ Human translators refine conceptual interpretation, adjust stylistic coherence, and ensure cultural 

accuracy. 

This hybrid model is particularly valuable for TCM classics because: 

1) AI accelerates workflow without compromising translator control. 

2) Humans correct conceptual errors and integrate cultural knowledge that AI cannot infer. 

3) The model supports broader dissemination: faster production + better cultural fidelity. 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

90 
Published by SCHOLINK INC. 

Such collaboration aligns with recent scholarship on “augmented translation,” which views AI as an 

assistant rather than a replacement for human translators. 

5.5 Implications for the Global Dissemination of TCM Knowledge 

Accurate translation of the Huangdi Neijing plays a critical role in shaping global perceptions of TCM. 

Misinterpretations—whether generated by AI or human translators—may distort core medical concepts, 

mislead readers, or reduce the text to a series of biomedical analogies. The findings underscore several 

implications: 

⚫ Translation accuracy is essential for preserving TCM epistemology. 

⚫ Cross-cultural medical education requires reliable translations grounded in classical logic. 

⚫ The growing presence of AI tools among TCM students and practitioners necessitates clear 

guidelines for their use. 

As the Spanish-speaking world continues to adopt TCM in clinical, educational, and popular contexts, 

translation quality will directly influence its integration and cultural legitimacy. 

5.6 Future Directions for AI-Assisted Translation of Classical Texts 

The limitations identified in this study also point toward potential avenues for future development: 

1) Domain-specific fine-tuning of AI models with TCM corpora may improve conceptual accuracy. 

2) Interactive human–AI workflows can reduce the risk of semantic drift. 

3) Standardized bilingual TCM terminology databases will enhance consistency across translations. 

4) Explainable AI systems may support transparency in interpretive processes. 

However, until AI systems acquire deeper conceptual modeling capabilities, human translators remain 

irreplaceable custodians of classical medical knowledge. 

 

6. Conclusion 

This study compared human and AI-generated Spanish translations of selected passages from the 

Huangdi Neijing to assess their performance in conveying the linguistic, conceptual, and cultural 

characteristics of this foundational TCM classic. The findings show that although AI translation tools 

have achieved notable advances in handling general linguistic structures and surface-level semantics, 

they remain insufficient for the accurate transmission of classical medical knowledge. Human translators, 

supported by domain expertise and cultural contextualization, demonstrated stronger interpretive 

capacity, especially in handling polysemous terms, conceptual networks, and cosmological frameworks 

central to TCM theory. 

The results reveal that AI excels in efficiency, consistency, and grammatical correctness, but lacks the 

depth required for translating culturally embedded concepts and philosophical-medical constructs. 

Human translators, while slower and occasionally inconsistent in formal accuracy, provide indispensable 

interpretive insight and cultural fidelity. Taken together, these findings suggest that neither approach 

alone can meet the full demands of translating the Huangdi Neijing and other TCM classics. 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

91 
Published by SCHOLINK INC. 

A key contribution of this study is the demonstration that a hybrid human–AI translation model is the 

most promising pathway for future translation work in this field. AI systems can support preliminary 

drafting, structural analysis, and terminological consistency, whereas human translators remain essential 

for accurate conceptual interpretation and cultural mediation. Such a model balances efficiency with 

fidelity, offering a practical solution for improving the accessibility and global dissemination of classical 

TCM knowledge. 

As interest in TCM continues to grow in Spanish-speaking regions, the quality of translation will directly 

influence its reception and integration into medical, educational, and cultural contexts. Future research 

should explore domain-specific AI training, collaborative workflows, and the creation of standardized 

bilingual terminology resources to further enhance translation quality. Ultimately, the findings reaffirm 

that while AI can facilitate translation processes, human expertise remains central to preserving the 

epistemological integrity and cultural depth of TCM classics.  

 

Declaration of AI Use 

The author used AI-assisted tools (ChatGPT) only for language polishing and structural adjustment 

during manuscript preparation. All conceptualization, data selection, translation analysis, and 

argumentation were conducted independently by the author. 

 

Funding 

This study was supported by  

1. the 2023 Research Project on International Chinese Language Education, titled “The Spanish 

Translation and Overseas Dissemination of the TCM Classic Huangdi Neijing” (Project No. 23YH67D)  

2. “Chinese Medicine Culture Promotion Center of Escuela Superior de MTC and Yunnan University 

of Chinese Medicine” 

 

Reference 

Castilho, S. (2023). Machine translation for the humanities: Opportunities and limitations. Digital 

Scholarship in the Humanities, 38(3), 612-628. 

Graham, A. C. (1986). Classical Chinese: A functional language perspective. Journal of Chinese 

Linguistics, 14(2), 115-149. 

Hightower, J. (1965). Problems of translating classical Chinese. Harvard Journal of Asiatic Studies, 

25(1), 92-102. 

Hurtado Albir, A. (2015). Traducción y Traductología: Introducción a la Traductología. Cátedra. 

Koehn, P. (2020). Neural Machine Translation. Cambridge University Press. 

https://doi.org/10.1017/9781108608480 

Lo, V. (2005). The influence of the Huangdi Neijing on East Asian medical traditions. Asian Medicine, 

1(1), 31-66. 



www.scholink.org/ojs/index.php/elsr              Education, Language and Sociology Research              Vol. 6, No. 4, 2025 

92 
Published by SCHOLINK INC. 

Munday, J. (2016). Introducing Translation Studies: Theories and Applications (4th ed.). Routledge. 

https://doi.org/10.4324/9781315691862 

Nida, E. A. (2001). Context and Meaning in Translation. Shanghai Foreign Language Education Press.  

OpenAI. (2024). GPT-4 Technical Report. Retrieved from https://openai.com/research 

Reiss, K., & Vermeer, H. J. (2013). Towards a General Theory of Translational Action. Routledge. 

https://doi.org/10.4324/9781315759715 

Sivin, N. (1987). Traditional medicine in contemporary China. The Journal of Asian Studies, 46(2), 291-

313. 

Unschuld, P. U. (2003). Huang Di Nei Jing Su Wen: Nature, Knowledge, Imagery in an Ancient Chinese 

Medical Text. University of California Press. 

ttps://doi.org/10.1525/california/9780520233225.001.0001 

Unschuld, P. U., & Tessenow, H. (2011). Huang Di Nei Jing Ling Shu: The Ancient Classic on Needle 

Therapy. University of California Press. 

Venuti, L. (2012). The Translator’s Invisibility: A History of Translation. Routledge. 

https://doi.org/10.4324/9780203553190 

 


