International Open Access Journal Platform

1 Introduction

In the contemporary era, the rapid advancement of deep learning and artificial intelligence technologies has revolutionized the landscape of language services, with machine translation emerging as a powerful and increasingly prevalent tool. Among the myriad of AI-driven translation platforms, DeepSeek and ChatGPT stand out as two of the most prominent contenders, garnering significant attention for their sophisticated natural language processing capabilities. Despite their shared goal of facilitating cross-linguistic communication, these two tools employ distinct technological approaches and are tailored for different application scenarios. As a result, their translation outcomes can vary significantly, necessitating a deeper exploration and comparative analysis to understand their respective strengths and weaknesses.

Guoqing Temple: The Cultural Treasure of Tiantai Mountain serves as an ideal case study for this purpose. This article is not merely a piece of text; it is a rich repository of historical information and Buddhist cultural elements, reflecting the intricate interplay between architecture, religion, and cultural heritage. Given its complex and culturally nuanced content, it provides challenging yet valuable material for comparing the translation differences between DeepSeek and ChatGPT. This study, therefore, aims to delve into the similarities and differences between these two AI translation tools in terms of language expression, cultural reproduction, and text creativity. It seeks to examine how effectively each tool can convey the subtle nuances of cultural texts while also exploring the reproducibility and creativity of machine translation in this context.

By conducting this comparative analysis, this research not only enriches the empirical findings of machine translation studies but also offers valuable references for translation practice. It sheds light on the potential and limitations of AI translation tools in handling culturally rich texts, thereby promoting a deeper integration of translation technology with the humanities. Ultimately, this study aspires to contribute to a more nuanced understanding of how artificial intelligence can be harnessed to bridge linguistic and cultural divides, paving the way for more effective and culturally sensitive translation practices in the future.

2 Text Selection and Evaluation Criteria

This section provides an overview of the research design, focusing on the selection of the source text and the establishment of evaluation criteria.

2.1 Text Selection

Guoqing Temple: The Cultural Treasure of Tiantai Mountain focuses on the cultural heritage of Tiantai Mountain and the history of Guoqing Temple, characterized by its strong cultural and historical background. According to Reiss’s (2000) theory, the text in this study is an informative text, emphasizing objective facts and clear expression. The analysis targets two categories: translations of four-character headings and complex sentences. Four-character headings in Chinese are concise and condensed, often carrying unique cultural connotations and rhetorical effects. Complex sentences, on the other hand, involve more profound historical backgrounds and cultural descriptions, testing the translation tools’ abilities in structural handling, semantic reproduction, and cultural transmission.

2.2 Evaluation Criteria

Li Changshuan (2025: 13), drawing on his translation practice and analyzing various discourses on translation standards and methods, proposed three standards—faithfulness, smoothness, and usability (corresponding to “faithfulness, expressiveness, and elegance”)—and their implementation methods: understanding, expression, and adaptation. These standards represent the common ground of translation theories across time and cultures. Based on this, the evaluation criteria for machine translation in this study are further divided into accuracy of translation (“faithfulness”), fluency of translation (smoothness or “expressiveness”), and creativity of translation (usability or “elegance”).

2.2.1 Accuracy

The accuracy of translation is the foundation for measuring translation quality. In terms of terminology selection, cultural texts often contain numerous proper nouns, historical terms, and specialized vocabulary. Accurate translation of these terms is crucial for conveying the cultural connotations of the original text. DeepSeek and ChatGPT may show
significant differences in the translation of terms such as “天台宗” and “隋塔”, which directly affect the cultural accuracy
and readers’ understanding of the translation. Additionally, the accuracy of semantic transmission is a key point of assessment. Translation is not only the conversion of words but also the reproduction of meaning. This study will compare the semantic correspondence between the original and translated texts to analyze the performance of the two translations in terms of information completeness and semantic accuracy, determining whether there is semantic deviation or information omission.

2.2.2 Fluency

The fluency of translation directly affects its readability and naturalness. Grammatical correctness is the basis for fluency. This study will examine whether the two translations contain grammatical errors or unreasonable structures. On this basis, the fluency and naturalness of sentence structure are also important considerations. Cultural texts often contain complex sentences, and translators need to adjust sentence structures while remaining faithful to the original text to adapt to the expression habits of the target readers. DeepSeek and ChatGPT may adopt different strategies when dealing with complex sentences, such as breaking down or reorganizing sentence structures to improve readability. This study will assess the performance of the two translations in terms of sentence structure fluency and naturalness, exploring whether they can present the original information in a natural and fluent manner without affecting the reading experience due to rigid sentence structures.

2.2.3 Creativity

Cultural adaptation and emotional expression are important manifestations of creativity. Cultural texts not only convey information but also carry rich cultural connotations and emotional colors. In terms of cultural background reproduction and handling, translation needs to remain faithful to the original text while taking into account the cultural background of the target language to avoid cultural misunderstandings or conflicts. For example, when translating the name Guoqing Temple, which has profound cultural connotations, it is necessary to consider how to convey its cultural significance in the target language. Additionally, emotional transmission is also an important aspect of creativity. Cultural texts often contain the author’s emotional inclinations and cultural attitudes, which need to be conveyed through expression. This study will analyze the performance of the two translations in emotional transmission and cultural context handling, exploring whether they can accurately convey the original emotional intentions while avoiding deviations due to cultural differences.

3 Translation Comparison and Quality Evaluation

This section selects four-character headings and complex sentences from Guoqing Temple: The Cultural Treasure of Tiantai Mountain as samples and compares the translation performance of DeepSeek and ChatGPT under the same instructions in terms of accuracy, fluency, and creativity. In addition to the required translations, DeepSeek provides a detailed textual thinking process, while ChatGPT includes a brief evaluation of the provided translation, facilitating subsequent editing. This confirms that generative intelligent translation has changed the interaction between translators and intelligent technologies. The translation process is no longer limited to static input and output but has become a dynamic, continuously improving collaborative translation process (Yang, Chen & Wei, 2025).

3.1 Lexical Level

This subsection focuses on the lexical level of translation, examining the choices and nuances in word selection made by DeepSeek and ChatGPT. At this level, the analysis explores how the two translation tools handle culturally specific terms, historical references, and Buddhist terminology found in the source text. By comparing their lexical choices, this section aims to reveal the strengths and weaknesses of each tool in conveying the precise meaning and cultural connotations of the original text. The findings will provide insights into the accuracy, appropriateness, and cultural sensitivity of their translations at the word level.

Example 1：草木皆佛

DeepSeek：Buddha in Nature/Sacred Flora

ChatGPT：All Things are Buddha

“草木皆佛” is a heading in the second section, not only depicting the tranquility and harmony of the natural scenery
in Guoqing Temple but also echoing the introduction of ancient plants such as the thousand-year-old plum tree from the Sui Dynasty. It also implies a philosophical idea that all things have a Buddha nature. When translating, considering that the following content includes not only plants but also stones, steles, and other scenic spots, the term “草木” does not only refer to plants but also represents all “sentient beings” other than humans; that is, the Zen concept of “insentient beings having Buddha-nature” (Zhang, 2010).

In its deep thinking process, DeepSeek specifically mentioned two points: First, the concept of all things having Buddha-nature may be unfamiliar in Western culture, so a concise title that conveys the connection between nature and Buddhism is needed. Second, considering the content of the following text, it is necessary to emphasize the history and spirituality of ancient trees. It provides a reference translation, “Ancient Trees, Living Buddhism”, but this was not adopted in the final result. The translation “Buddha in Nature / Sacred Flora” by DeepSeek better conveys the Buddha-nature in natural scenery, bringing a sense of tranquility and harmony. However, this translation is somewhat limited because it focuses mainly on plants and fails to fully cover the broader philosophical idea of “sentient beings” including stones and steles. On the other hand, ChatGPT’s translation “All Things are Buddha” is more accurate in conveying the philosophical idea that all things have a Buddha nature, aligning with the deeper meaning of the original text. In terms of fluency, “Sacred Flora” appears somewhat overly academic and may hinder reader comprehension. In terms of creativity, the broad sense of “All Things” in “All Things Are Buddha” better fits the context of the original text.

Example 2：知恩报恩

DeepSeek：Gratitude in Stone

ChatGPT：Gratitude and Repayment

On an important Buddhist relic, the Lotus Sutra Stupa in Guoqing Temple, the phrase “知恩报恩” is inscribed. The first character, “恩” is larger than the second, indicating that the favor received from others is greater than the favor one returns. It encourages people to always maintain a grateful heart and spread kindness.

Influenced by the provided context, DeepSeek translated “知恩报恩” as “Gratitude in Stone”, intending to convey
the nature of the inscription being carved in stone. However, it failed to fully express the complete meaning of “报恩” (repayment), shifting the focus to the material form of “stone”, which differs from the original emphasis on the interactive and reciprocal relationship. During the deep thinking process, DeepSeek emphasized the need for a concise translation that conveys both gratitude and repayment, considering the textual form and providing reference translations such as “Gratitude and Giving Back” or “Grace and Gratitude”, which are not only more accurate but also use alliteration, demonstrating greater creativity. However, DeepSeek did not adopt these translations in the final result.

In comparison, ChatGPT’s translation “Gratitude and Repayment” is more accurate, clearly expressing both aspects of “知恩” (gratitude) and “报恩” (repayment) and reflecting the relationship between the favor received and the favor returned in the original text.

In terms of fluency, DeepSeek’s translation is concise but appears somewhat awkward and confusing, making it difficult for readers to understand its intended meaning. In contrast, ChatGPT’s translation is straightforward and easy to understand.

Regarding creativity, DeepSeek’s “Gratitude in Stone” has a certain uniqueness, giving a sense of enduring gratitude through the association with stone. However, it lacks cultural and philosophical depth due to the incomplete
expression of “报恩”. ChatGPT’s translation is more direct, lacking artistic and cultural symbolism, but it clearly conveys
the core theme of gratitude and repayment.

Example 3：夷夏之争

DeepSeek：Buddhism’s Cultural Crossroads/ The Great Buddhist Debate

ChatGPT：The Debate of Foreign and Native

“夷夏之争” refers to a historical debate about whether Buddhism should survive and develop in China, closely related to the core issues of cultural identity and religious conflict in the text. The term “夷” generally refers to ethnic groups outside the Huaxia (Han Chinese), and it was once a derogatory term for non-Han peoples in ancient China. When translating, it is necessary to maintain the sensitivity and adaptability required in cross-cultural communication.

In terms of accuracy, DeepSeek provided two translation options: Buddhism’s Cultural Crossroads and The Great Buddhist Debate. Both attempts to convey the cultural conflict and religious debate from different angles but failed to directly reflect the cultural opposition between “夷” (foreign) and “夏” (native). In contrast, ChatGPT’s translation “The Debate of Foreign and Native” more accurately captures the core meaning of the original text, with “Foreign” and “Native” corresponding to “夷” and “夏”, respectively, and more precisely reflecting the debate on cultural conflict and identity.

In terms of fluency, DeepSeek’s translations, while conveying some cultural and religious content, appear somewhat cumbersome in structure. For example, “Buddhism’s Cultural Crossroads” feels too vague and does not provide an intuitive understanding for readers. In contrast, ChatGPT’s translation “The Debate of Foreign and Native” is concise and clear, easy to understand, and effectively expresses the core meaning of the original text, especially in contrast between “Foreign” and “Native”, which directly highlights the theme of cultural opposition.

Regarding creativity, ChatGPT’s direct translation avoids using “Barbarians” for “夷”, demonstrating sensitivity and adaptability in cross-cultural communication. DeepSeek’s translations, through terms like “Cultural Crossroads” and “Great Debate”, attempt to create an atmosphere of cultural collision and intellectual confrontation, showing a certain degree of creativity.

3.2 Syntax Level

This subsection delves into the syntax level of translation, analyzing how DeepSeek and ChatGPT structure sentences and convey meaning through grammatical arrangements. The focus is on their ability to maintain coherence, readability, and naturalness in the target language while accurately representing the source text’s syntactic complexity. By examining their approaches to sentence construction, this section aims to assess the fluency and grammatical accuracy of the translations. The analysis will highlight differences in syntactic strategies and their impact on the overall quality and comprehensibility of the translated text.

Example 4：唐会昌年间，佛教发展过于迅速，引起了朝廷的严重不满，唐武宗下诏灭佛，国清寺遭到了空前的劫难。直至唐宣宗即位，下诏重修寺宇，国清寺才得以在废墟中重建。

DeepSeek：In the Tang Dynasty, the development of Buddhism became too swift, leading to significant dissatisfaction among the government. Emperor Taizong issued an edict to eradicate all Buddhists, resulting in the great劫难 of the Qiqi Temple. It was not until Emperor Xuanzong, the Tang’s sixth emperor, issued an edict to rebuild the temple that Qiqi Temple was repaired in ruins.

ChatGPT：During the reign of Emperor Wu of the Tang Dynasty, known as the Huichang period, the rapid spread of Buddhism provoked severe dissatisfaction within the imperial court. In response, Emperor Wu issued an edict to suppress Buddhism, leading to an unprecedented disaster for the Guoqing Temple. It was not until the reign of Emperor Xuanzong, several decades later, that the temple was ordered to be rebuilt, allowing Guoqing Temple to rise from the ruins.

DeepSeek’s translation generally conveys the main idea of the original text but contains several inaccuracies. First, the original text mentions“唐会昌年间”, but DeepSeek translates it as “in the Tang Dynasty”, lacking the specific time frame. The Huichang period was a specific era during Emperor Wu’s reign, which should be translated as “the Huichang
period of the Tang Dynasty”. Additionally, “唐武宗” is mistakenly translated as “Emperor Taizong”. Emperor Wu (Li Yan)
and Emperor Taizong (Li Shimin) are different monarchs, resulting in a clear error. The original “国清寺” is mistranslated
as “Qiqi Temple”. Finally, “劫难” is translated as “great 劫难” which is unclear and should be directly translated as “great calamity” or “disaster”. ChatGPT’s translation is more accurate, clearly mentioning “Emperor Wu” and correctly using “Emperor Wu of the Tang Dynasty”. The original “会昌年间” is accurately translated as “the Huichang period”.
Additionally, “国清寺” is correctly translated, and “劫难” is accurately expressed as “unprecedented disaster”. Overall,
ChatGPT performs better in terms of accuracy.

In terms of fluency, although DeepSeek’s translation has a clear sentence structure, some expressions are too direct or unnatural. For example, “the great劫难 of Qiqi Temple” with “great 劫难” transliterated from Chinese does not integrate well with the English context, and the mistranslation of “Qiqi Temple” also affects the fluency of the sentence. Additionally, “that Qiqi Temple was repaired in ruins” appears somewhat abrupt and could be more smoothly expressed as “that Guoqing Temple was rebuilt from the ruins”. ChatGPT’s translation is more natural and elegant in fluency. For example, “the rapid spread of Buddhism provoked severe dissatisfaction within the imperial court” is precise and smooth, with a clear, logical structure and natural expression. Additionally, the latter part, “allowing Guoqing Temple to rise from the ruins,” is smooth and visually impactful, making it very readable. Overall, ChatGPT’s translation is more fluent.

In terms of creativity, DeepSeek’s translation is relatively straightforward and lacks significant creative expression. For example, “great劫难” is quite direct and lacks a vivid expression. The overall sentence structure is plain, conveying information accurately but conservatively in terms of creativity. In contrast, ChatGPT’s translation is more creative, especially in the expression “rise from the ruins”, which adds more imagery and vividness to the passage. Additionally, the use of “the rapid spread of Buddhism” and “unprecedented disaster” adds more depth and expressiveness to the translation, demonstrating higher creativity.

Example 5：《管子》指出：“畜之以道，则民和；养之以德，则民合。和合故能习；习故能偕。”

DeepSeek 1：According to The Book of Moments (Guanzi), “Educate people with virtue, and they will harmonize; Nurture them with virtue, and they will unite. Union hence enables learning; Learning hence enables harmony.” This passage emphasizes the importance of virtue in fostering social harmony.

ChatGPT 1：涉及敏感词，无法翻译

DeepSeek 2：The book Guanzi states: “Only by guiding the people through moral education can they adapt to socialized life, thereby achieving harmony and unity in society.”

ChatGPT 2：The book Guanzi points out: By guiding the people through moral education, they will be able to adapt to social life, and society can achieve harmony and unity.

Initially, ChatGPT deemed the sentence to involve sensitive terms and was unable to generate a translation. However, during the deep thinking process, DeepSeek accurately captured the need to first understand the meanings of terms such as “畜” , “养”, “和合”, and “习” in the sentence. It concluded that “畜” and “养” refer to “education”
and “cultivation”, while “和合” and “习” respectively mean “harmony and unity” and “learning or studying”. This understanding
is not merely dictionary-based but involves, as Li Changshuan (2025: 13) suggested, a deeper comprehension of the specific meanings in context and their relationships with related concepts. Here, “和合” means harmony and unity, and “习” does not mean “learning” but rather “habituation (to social life)”.

After translating the original text into modern Chinese as “《管子》一书指出：用道德教育引导人民，他们才能习惯社会化生活，社会才能和谐统一”, DeepSeek and ChatGPT generated their respective translations. DeepSeek correctly identified the book title and italicized it in the translation, while ChatGPT overlooked this detail.

In terms of accuracy, DeepSeek’s translation is more complete in conveying the meaning, using the structure “Only by... can...” to emphasize the necessity of moral education in guiding people to adapt to social life. It also accurately
translated“和谐统一”as “harmony and unity”, better reflecting the original text’s logical relationship and main point.
ChatGPT’s translation, while conveying the basic meaning of the original text, is somewhat flat in expression. It used the structure “By... they will be able to...” which, although logically clear, lacks the emphasis of the original text. Additionally, “social life” is slightly broader than DeepSeek’s “socialized life”, failing to precisely convey the
specific meaning of “社会化生活” in the original text. In terms of fluency, DeepSeek’s translation is smoother, with a
concise sentence structure that is easy to understand and conforms to English expression habits. In terms of creativity, DeepSeek’s translation, through the use of “Only by... can...”, enhances the emphasis of the sentence, making the translation more persuasive and demonstrating a deeper understanding of the original text’s logical relationship. Additionally, the choice of “socialized life” is closer to the original meaning, showing higher creativity. In contrast, ChatGPT’s translation focuses more on directly conveying the meaning of the original text. Although the language expression is natural and smooth, it lacks creativity in special structural or lexical choices to enhance the expressiveness of the translation.

Example 6：法华经幢是国清寺内一座重要的佛教文物，上书“知恩报恩”。由图可见，第一个“恩”字比第二个“恩”大，意为他人给予的恩情大于自己回报的恩情，鼓励人们常怀感恩之心，传递善意。

DeepSeek：The幢 (/Guangzi/) of the “Mandate of Rebirth” at Qiqi Temple is an important Buddhist relic. It bears the inscription “Know恩报恩” (“Recall恩典 and repay恩报”). The first “恩” is larger than the second, indicating that the恩典 (grace or favor) received is greater than the恩报 (remonstrance or return). This statute encourages pilgrims
to maintain a grateful heart and extend kindness to others.

ChatGPT：The Lotus Sutra Stupa is an important Buddhist artifact located within Guoqing Temple, inscribed with the phrase “Gratitude and Repayment”. As shown in the image, the first character for “恩” (favor or grace) is larger than the second, symbolizing that the kindness received from others is greater than the kindness one repays. This inscription encourages people to always maintain a heart of gratitude and to spread goodwill.

DeepSeek’s translation has several inaccuracies. First, “法华经幢” is mistranslated as “The幢 (/Guangzi/) of the
‘Mandate of Rebirth’”, which is influenced by the previous question and does not align with the original meaning. According to the context, “法华经幢” should be accurately translated as “Lotus Sutra Stupa” or “Stupa of the Lotus
Sutra”. Besides, “Mandate of Rebirth” is not the common translation of “法华经幢” and does not correspond with the religious background shown in the source text. Additionally, the translation of “恩”, such as “恩报” and “恩典”,
is not entirely accurate and does not closely align with the Chinese context. For example, “恩报” is translated as “remonstrance
or return”, which is not entirely appropriate; “repayment” would be more fitting to convey the original meaning. In contrast, ChatGPT’s translation is more accurate, closely adhering to the original text. First, “法华经幢” is correctly translated into “The Lotus Sutra Stupa”, which is a common translation for Buddhist relics and accurately conveys the original meaning. The term“恩” is appropriately translated as “favor or grace” and “kindness”, which are more in line with the context and meaning of the original text. Additionally, “恩报” is accurately translated as “repayment”, which fits the intended meaning. Overall, ChatGPT’s translation is more precise in terms of accuracy.

In terms of fluency, ChatGPT’s translation is more natural and smooth. It uses concise and clear language to convey the meaning of the original text, with a logical sentence structure that is easy to read and understand. For example, the sentence “As shown in the image, the first character for ‘恩’ (favor or grace) is larger than the second”, clearly explains the situation in the image and smoothly transitions to the subsequent explanation. In contrast, DeepSeek’s translation appears more rigid and verbose. For example, the sentence “It bears the inscriptions ‘Know恩报恩’(‘Recall恩典 and repay恩报’)” , is not only inaccurate in translation but also uses parenthetical notes to explain “恩典” and “恩报”, which make the expression less natural and affects the overall fluency.

In terms of creativity, ChatGPT’s translation is also more advantageous. It not only accurately conveys the meaning of the original text but also enhances the translation’s appeal and expressiveness through details such as translating “传递善意” as “spread goodwill”. In contrast, DeepSeek’s translation lacks this creativity, focusing more on a direct translation of the original text without adequately processing and refining the language, resulting in a lack of creativity.

4 Conclusion

Wang Huashu and Zhang Chengzhi (2025: 56) pointed out that machine translation can save time and reduce costs, while humans can review and improve the quality of machine translations. The complementary advantages of humans and machines (Dai & Liu, 2023: 86) will make human-machine collaboration a common practice in translation, with vertical domain intelligent translation on the rise.

DeepSeek demonstrated a profound understanding of cultural background and strong adaptability to context, especially in handling proper nouns and metaphorical expressions related to Buddhist culture. For example, in translating “夷夏之争”, DeepSeek’s translations of “Buddhism’s Cultural Crossroads” and “The Great Buddhist Debate” cleverly combined cultural conflict with religious debate, showing sensitivity to cultural background. Additionally, DeepSeek’s deep thinking process, which is more valuable than the generated results, provides rich references for translation practice through its logical reasoning and contextual analysis in translation decisions. However, ChatGPT excelled in language fluency and logical coherence, with concise and clear translations that are easy to understand. ChatGPT’s translation “The Debate of Foreign and Native” for “夷夏之争” directly conveyed the theme of the debate between “foreign” and “native”, with a more straightforward language expression. Additionally, ChatGPT’s self-evaluation function allows it to assess and optimize its translations, further enhancing the accuracy and reliability of the translations. During the usage process, individual translators should maintain the habit of independent thinking and an exploratory mindset, adopting and rigorously verifying the responses from both tools (Wang & Jiang, 2024).

While this study has provided valuable insights into the characteristics and differences of DeepSeek and ChatGPT in translating culturally rich texts, it is essential to acknowledge its limitations and areas for improvement. First, t the scope of the research is confined to a single source text, focusing only on the article Guoqing Temple: The Cultural Treasure of Tiantai Mountain. Although this text offers a wealth of cultural and historical context, the limited sample size restricts the generalizability of the findings. Future research could benefit from a more extensive range of cultural texts to provide a broader understanding of how these translation tools perform across diverse contexts. Second, the study has focused primarily on the outcomes of the translations without delving into the underlying technical aspects of the translation process, such as differences in model architecture and training data that may affect translation outcomes. Additionally, the study has not fully considered the reception of the translations by different audience groups. Translation effectiveness is not solely determined by linguistic and cultural accuracy but also by how well the translated content resonates with its intended audience. Future research could incorporate audience feedback to more comprehensively assess translation effectiveness.

Despite the convenience that technological advancements have brought to translation practice, machine translations often have shortcomings, especially in terms of humaneness. Therefore, post-editing has become an essential step. This requires translators to not only master new technologies but also possess a high sense of responsibility and humanistic thinking abilities. Human translators, with their unique humanistic values and creativity, will continue to play an irreplaceable role in the integration of technology and humanities, driving the translation industry toward a brighter future. Just as Ye Zinan (2023: 178) stated, As long as the sun continues to rise in the east and people continue to breathe, the “replacement theory” will not become a reality. In the field of translation, it is humans who will continue to lead the way.

Reference

1. Dai, G. R., & Liu, S. Q. (2023). Neural machine translation: Progress and challenges. Foreign Language Education, (1), 82-89.
2. Duan, Y. C. (2025, February 15). Seizing AI discourse power: DeepSeek’s technological advantages, strategic layout and future ecological vision. Journal of Xinjiang Normal University (Philosophy and Social Sciences Edition), (4), 1-10.
3. Fang, X. D., Wang, B., & Zhong, X. M. (2025, February 15). The DeepSeek moment: Technology-communication-society (TCS) framework and bridging the mainstreaming gap. Journal of Xinjiang Normal University (Philosophy and Social Sciences Edition), (4), 1-10.
4. Huang, S. X., & Huang, C. (2023, August 16). Governance wisdom in Guanzi [OL]. Qinfeng Network. Retrieved February 13, 2025, from https://wap.qinfeng.gov.cn/info/1064/217070.htm
5. Li, C. S. (2025). Bridging Chinese and foreign translation theories and practices with “faithfulness, expressiveness and closeness”. Shanghai Journal of Translators, (1), 13-18.
6. Reiss, K. (2004). Translation criticism: The potentials and limitations. Shanghai Foreign Language Education Press.
7. Sun, J. (2016). Machine translation research based on text typology theory. Chinese Science & Technology Translators Journal, 29(3), 27-32.
8. Wang, H. S., & Zhang, C. Z. (2025). Translation practice models in the GenAI era: Technological iteration, industrial transformation and future trends. Foreign Language Education, (1), 53-58.
9. Wang, J. H., Jiang, D. C., & Song, J. N. (2024). ChatGPT and quantitative translation evaluation. Technology Enhanced Foreign Language Education, (6), 80.
10. Wang, J. S., Zhuang, C. Q., & Wei, Y. P. (2024). Machine translation quality assessment: Methods, applications and prospects. Foreign Language and Literature, (3), 135-144.
11. Yang, Y. X., Chen, Y., & Wei, X. Q. (2025). Research on human-machine collaborative translation literacy in the generative AI era. Shanghai Journal of Translators, (1), 39-45.
12. Ye, Z. N. (2023). In the ChatGPT era, human translation remains unparalleled. Chinese Translators Journal, 44(3), 173-178.
13. Yuan, X. Y. (2025). “Subjectivity” and “creativity” in AI literary translation. Journal of Shanghai Jiao Tong University (Philosophy and Social Sciences), (1), 1-10.
14. Zhang, A. L. (2010). The theory of “non-sentient beings having Buddha-nature” in Chinese Buddhism. Journal of Southwest University (Social Sciences Edition), 36(6), 81-84.
15. Zhang, H. M. (2025, February 11). How was DeepSeek-R1 developed? Journal of Shenzhen University (Science and Engineering Edition), 1-7. https://link.cnki.net/urlid/44.1401.N.20250210.1628.002