Both cases use Chinese source texts and English translations, but the two genres ask very different things of the translator.
Academic research · Translation technology · 2026
Chinese to English with AI assistance: news vs literature
What changes when the source genre changes?
This academic study compares two Chinese-to-English translation tasks: a contemporary CCTV news article and an excerpt from Tao Yuanming’s The Peach Blossom Spring. I used AI to support initial understanding and drafting, then revised the English myself and documented where the process was useful, where it failed and where specialist review would still be required.
This was an exploratory study, not a claim of professional Chinese competence.
AI supported initial understanding and drafting. I revised the English and made the final wording decisions.
For the news case, I recreated the article layout on a website with Chinese and English views.
A native Chinese speaker read the news translation and helped check relevant meaning errors. This was an external reading, not a professional review.
Scope and limitations
What I could and could not claim.
I do not have the independent command of Chinese required for professional Chinese translation. In this study, I worked through AI assistance, research and external checks. English was the language I could properly assess and revise.
The news translation gave me more opportunities to verify structure, facts and terminology. The literary translation remained much more dependent on interpretation and would require specialist review before professional use.
01 / Research question
Same language pair, very different texts.
I wanted to see how much the task changed when the source moved from contemporary news to classical literature while the language direction and use of AI stayed the same. My expectation was that an informational text would be easier to verify, while literature would expose more problems around ambiguity, voice, rhythm and interpretation.
02 / Method
How I approached it.
The study used two case studies, AI-assisted drafting, human revision and comparison. For each text, I looked at its purpose and structure, the linguistic and cultural difficulty, the amount of rewriting needed, the demands on English-language judgement and the risk of errors I might not be able to detect.
News case
A CCTV article about agricultural technology, water saving, mechanisation and grain production. The English version retained the original titles, paragraphs, images and reading structure.
Literary case
An excerpt from 桃花源记, attributed to Tao Yuanming. Its concision, ambiguity, cultural imagery and rhythm made the task much less reducible to literal meaning.
AI role
AI helped with initial understanding and drafting. I then reviewed the output for meaning, register, naturalness, units, presentation and literary effect.
Evaluation
I separated decisions I could check against sources from decisions that depended more heavily on interpretation, then documented what still needed specialist review.
03 / News case
Where the news article was easier to verify.
The CCTV article had a familiar news structure and clearly stated facts. AI helped me get to a first draft, but the English still needed significant revision. I focused on literal phrasing, register, Chinese units of measurement and normal English news prose.
Chinese units that many English readers would not recognise. I converted or explained them where that improved comprehension.
I rendered this compact technical phrase as “plastic film mulching, precision hole sowing and drip irrigation” to make the agricultural process explicit.
I reworked sentence structure and repetition where the first draft sounded too close to the Chinese.
The companion website keeps the article hierarchy and lets readers switch between Chinese and English.
04 / Literary case
Where interpretation mattered more.
The Peach Blossom Spring required decisions about concision, ambiguity, rhythm, imagery and atmosphere. AI could suggest approximate meanings, but it could not decide what kind of English literary experience the translation should create.
I used “it suddenly opened out before him” to convey the sudden opening of the landscape.
I used “the old and the young”. The paper notes the trade-off: the English is clearer, but some of the original imagery used to evoke age is lost.
The English needed to carry the movement from a narrow passage into a broad landscape, not just reproduce individual meanings.
An analytical translation proposal that would require specialist review before professional use.
05 / Comparison
What changed between the two.
| Dimension | News article | Literary excerpt |
|---|---|---|
| Function | Communicate factual information clearly. | Create atmosphere, imagery and a literary experience. |
| Main quality criteria | Accuracy, readability and terminology consistency. | Voice, rhythm, ambiguity and cultural density. |
| Main risk | Factual errors or unclear units of measurement. | Loss of style or over-smoothing the original. |
| Role of AI | Useful support for understanding and drafting in this case. | More limited support, with greater demands on human interpretation. |
06 / Professional responsibility
The final language still has to be assessable.
The study argues that a tool can help someone approach a source language they do not fully command, but it does not remove the responsibility to assess the final text. I could explore Chinese-to-English because I could judge and revise the English. I could not responsibly offer the reverse direction as professional work.
07 / Conclusion
What I took from the exercise.
AI was useful for getting started, especially with the news article. It was much less useful for deciding whether the final English had the right register, rhythm or literary effect. The main lesson for me was that the value of a translation still depends on decisions someone can understand, justify and take responsibility for.
Paper and companion website
Read the study or open the translated article.
The paper includes the method, comparison, limitations, references, source excerpts and translation decisions. The companion website recreates the CCTV article with Chinese and English views.
Download the full paper
14 pages including analysis, references and appendices.
Download the full paperOpen the companion website
A recreation of the CCTV article with Chinese and English views.
Open the companion websiteTranslation expertise