Translation usually happens at the very end.

The partnership agreement has been signed. The CPD programme has been designed and all the training materials are ready. Then comes the inevitable question: "Can someone translate this into Chinese?"

Someone opens an AI translator.

Five minutes later, the document is finished.

Or so they think.

Modern translation tools are remarkably good, but universities aren't translating everyday conversations—they're translating academic systems, professional terminology and institutional culture. That's where things start to become interesting.

Sometimes, very interesting.

Take this example. 
 

A UK university refers to its whole CPD programme. The Chinese translation becomes 培训教材 (péixùn jiàocái). 

Literally, that means just "training materials".

A Chinese reader is likely to think they're receiving a folder of teaching materials rather than enrolling on a structured professional development programme. 

A more appropriate translation might be 持续专业发展项目 (chíxù zhuānyè fāzhǎn xiàngmù) or simply CPD 项目 (CPD xiàngmù) if the audience is already familiar with the term. 

Nobody has made a grammatical mistake. They've simply translated the phrase word for word instead of conveying its intended meaning. 

Another common translation error involves the word School. 

A UK university proudly introduces its School of Engineering. 

A literal translation becomes: 工程学校 (gōngchéng xuéxiào)

To a Chinese reader, this simply means an Engineering school. 

In a higher education context, the more appropriate translation is: 工程学院 (gōngchéng xuéyuàn). 

The difference looks tiny, but it represents a completely different organisational structure.

Then there's the phrase every university loves: 

Flexible Learning. AI translation usually produces: 灵活学习 (línghuó xuéxí). 

Every Chinese speaker understands those words. The problem is that it sounds in Chinese as though students are free to study whenever and however they like—turn up on Tuesday, skip Wednesday, and perhaps finish the course next month if they feel inspired. 

A university is actually referring to flexible modes of delivery, such as online learning, blended learning or part-time study. A more natural translation might be: 灵活学习模式 (línghuó xuéxí móshì) or 弹性学习方式 (tánxìng xuéxí fāngshì), depending on the context. 

Just one or two extra words turn a literal translation into an accurate one.

Our personal favourite comes from almost every UK university website. 

"Our students are challenged to think critically." 

This makes perfect sense in English. 

Machine translation often translates as: 我们挑战学生批判性思考 (wǒmen tiǎozhàn xuéshēng pīpànxìng sīkǎo). 

It isn't grammatically wrong.

It sounds as though the university is picking a fight with its students and far more aggressive than intended. 

A Chinese academic would be more likely to write: 我们鼓励学生培养批判性思维 (wǒmen gǔlì xuéshēng péiyǎng pīpànxìng sīwéi). "We encourage students to develop critical thinking". 

Good translation sometimes requires rewriting the sentence, not translating it.

It's about conveying the intended meaning within the right context.

Universities use words like programme, school, faculty, module, assessment and credit every day. Those words seem ordinary until they cross borders. Then each one carries assumptions based on entirely different education systems.

AI can translate vocabulary. It cannot always determine the intended meaning behind that vocabulary.

So how do universities avoid these problems?

Fortunately, they can reduce these risks by following a few good practices.

• Choose translation professionals with first-hand experience of both countries who understand your audience. They do more than translate words. They adapt the language, tone and terminology to suit the intended readers, drawing on an understanding of how people communicate and interpret information in different educational, social and professional contexts.

• Build a shared terminology list. Agree on standard translations with all project partners for key terms such as School, Faculty, Programme, Module and CPD before translating hundreds of pages. 

• Always include a final human review. AI is an excellent starting point, but it shouldn't have the final say on documents representing your institution. 

Good translation should also be invisible. If readers notice the translation, it's probably because something doesn't sound quite right.

Universities spend decades building their academic reputations. A carefully written document shouldn't lose its meaning simply because it crossed a language boundary.

(This piece is intended as informal commentary to encourage discussion, not as a formal statement of policy.)