Gains:
- Ability to adapt content with artificial intelligence according to literacy level, language and cultural context
- Understand that health translation must be verified by a native speaker and that artificial intelligence translation cannot be trusted alone in critical messages.
- Knowing that artificial intelligence writes according to the average and being able to check local example, cultural assumptions and visual suitability as a midwife
Adapting Content for Cultural Responsiveness, Multilingualism, and Low Literacy
A midwife in Türkiye does not work with a single "average pregnant woman". You will encounter expectant mothers who speak different languages, have different literacy levels, and come from different beliefs, traditions and family structures. A message that works for one pregnant woman may seem meaningless or even hurtful to another. Cultural sensitivity—adapting care and communication while respecting the person's language, values, and life context—directly impacts midwifery outcome; A suggestion that is not understood will not be implemented, a hurtful message will break trust. Artificial intelligence is a powerful aid in this adaptation: it translates a text into different languages, reduces it to a literacy level, rewrites it according to different cultural contexts. But there is a critical limit here too: AI writes according to the average; The midwife (and, if possible, a native speaker) always checks for local and cultural accuracy.
In this unit you will learn how to adapt content for multilingualism, low literacy and cultural diversity; We cover where AI is helpful and where it can be misleading.
Caution: AI translation may be fluent, but it may mistranslate a medical term, miss a cultural reference, or use a non-native expression. A health-related translation must be reviewed by someone who knows the language — ideally a professional translator or a native-speaking healthcare professional.
Three dimensions of adaptation
1. Language (translation): Translating the text into the pregnant woman's native language. AI produces first translation quickly; but human control is essential for medical accuracy and natural expression. Incorrect translation causes harm, especially in urgent/critical messages.
2. Literacy level: Simplifying the same information so that a mother with limited literacy can understand it. Short sentences, daily words, very few terms, structure that can be supported with pictures. AI will quickly reduce a text to a “6th grade” or “very plain” level.
3. Cultural context: Rewriting nutritional examples to local cuisine, family role distribution, respecting real life and religious/traditional practices. AI gives an outline; The midwife knows which example is appropriate and what may be hurtful.
Why is health literacy so important?
Health literacy is a person's ability to find, understand and use health-related information. A mother with low health literacy may not be able to apply what is in it even if she receives the right brochure. This is not a matter of intelligence, it is a matter of communication not fitting the person. The task of midwifery is not to "give" information, but to make it understandable and applicable. AI makes this job easier by adapting the same content to different levels.
Step by step: adapting a message across multiple channels
- Clarify the source text: First, have a correct, verified Turkish text.
- Produce literacy versions: Ask the AI for two versions, “standard” and “very basic.”
- Translate to language: Have the AI do the initial translation into the required language(s).
- Human check: Review the translation with a native speaker (translator/colleague); Check medical terms specifically.
- Cultural adaptation: Put examples, references into local context; Clear out any hurtful/inappropriate items.
- Back translation test (optional): Translate the translation back to Turkish with another tool to roughly check that the meaning is preserved — but this is not a substitute for human checking.
Of these six steps, the fourth one is most often skipped: human control. The AI translation appears so fluid that a midwife who doesn't know the language will assume the text is "good." However, fluency is not proof of accuracy; A translation may be grammatically perfect but medically inaccurate. So never release the translation without knowing the language. Until a native-speaking colleague, an in-house translator, or a reliable translation service is available, delivering critical health messages solely through AI translation leaves the pregnant woman open to misunderstanding and harm.
three mini cases
Case 1 — Medical error in translation: A midwife has the instructions for "glucose test on an empty stomach" translated into another language by AI. When a native-speaking colleague reads the translation, he realizes that the phrase "on an empty stomach" has misunderstood the meaning of "having a light meal" in that language—which would have distorted the test result. It's being fixed. The AI was fast, but human control prevented damage.
Case 2 — Literacy adaptation: A second midwife notices that the brochure she gave to a group of pregnant women with limited literacy is not being implemented. He asks AI to rewrite the brochure in a "very simple way, each sentence containing no more than 8 words, supported by a picture" and adds visuals. In the next period, the rate of implementation of the recommendations increases significantly. The information has not changed; It just became understandable.
Case 3 — Cultural sample fit: In the third example, the AI includes samples not commonly found in the local cuisine in a nutrition brochure. The midwife replaces these with foods that are common in the region, affordable and culturally familiar. It also relaxes an assumption the AI makes about sharing decisions in the family (which is not the same in every family). The brochure becomes both accurate and “ours.”
Copiable templates
Role: You are the literacy adaptation assistant.Task: Give the following text in TWO versions: (A) 6th grade level, (B) very simple (each sentence no more than 8 words, supported by pictures, no terms).CHANGE meaning, ADD/REMOVE information.Text: [text]
Role: You are the initial translation assistant (NOT the final translation). Task: Translate the following healthcare text into [target language]. Rules: Mark medical terms ("[TERM - HUMAN CONTROL]"). If you are not sure, do not translate, leaving "[UNCERTAIN]". This translation will be verified by a human; be careful accordingly.Text: [text]
Role: You are a cultural adaptation consultant.Task: Review the examples, references, and assumptions in the text below. Check items that MAY NOT fit the following region/culture context:[region/culture information]. Suggest a more suitable alternative for each.Rules: DO NOT CHANGE clinical accuracy; just increase cultural fit.Text: [text]
Role: You are the back-translation control assistant. Task: Translate the following [target language] text back into Turkish. Just give the translation; I will compare it with the original and check the semantic shift myself.Text: [translation]
Weak prompt / Strong prompt
Weak: "Translate this brochure into Arabic."
Result: A translation that seems fluent, but whose medical terms are not checked, whose cultural compatibility is unclear, and which cannot be trusted on its own.
Strong:
Role: first translation assistant. Translate this health text into [target language]. Mark medical terms with "[TERM - HUMAN CONTROL]" and leave "[UNCERTAIN]" where you are not sure. This translation will be verified by a translator/native speaking healthcare professional.
Difference: The powerful prompt positions the translation as a "first draft" rather than a "final" one, flags risky areas, and requires human control.
Adaptation size and responsibility
Size
AI contribution
human control
Literacy download
Quick two versions
Midwife tests language
Translation
first draft
Translator/native speaker
Cultural example/reference
Alternative suggestion
Midwife knows local availability
Religious/traditional sensitivity
general outline
Midwife/community information
visual harmony
Image note
Midwife clinical/cultural accuracy
Tip: Never rely on AI translation alone for critical or urgent messages (e.g. "apply now" indications). The cost of incorrect translation in such messages is very high; Professional translation or native speaking healthcare professional control is mandatory.
Common mistakes
- Thinking that AI translation is final: Health translation must be subject to human control.
- Not specifying literacy level: If you do not specify, the AI will write at an average-high level.
- Ignoring cultural assumptions: AI can make non-native assumptions about family/role/nutrition.
- Leaving non-local samples: Food/practice recommendations that are not available in the region are not applied.
- Not checking the visual culturally: The visual may also contain inappropriate/foreign elements.
- Thinking that one language and one level is enough for everyone: Every pregnant woman's language and literacy are different.
In summary
Cultural sensitivity and accessibility directly determine the outcome of care in midwifery: a recommendation that is not understood will not be implemented. Artificial intelligence accelerates this task by translating content into multiple languages, reducing it to a literacy level, and adapting it to cultural context. But AI writes according to the average; The translation must be verified by a native speaker, and the midwife must check cultural compatibility and local examples. AI translation alone cannot be relied upon for critical messages. The aim is not to "give" information, but to make it understandable and applicable.
Application task
Get a verified Turkish study material. Make two versions with the literacy adaptation template and test the “very simple” version with a limited reader. Then have the material translated into a target language and, if possible, review it with a native-speaking colleague to find at least one translation/cultural error. Finally, replace a non-native example with an example appropriate for your region with the cultural adaptation template.
checklist
- [ ] I first verified the accuracy of the source text in Turkish.
- [ ] I typed the literacy level into the prompt and tested the plain version.
- [ ] I received the translation as a "first draft" and had it human-checked.
- [ ] I provided professional/native language control on critical/urgent messages.
- [ ] I adapted cultural examples and assumptions to the local context.
- [ ] I checked the cultural and clinical appropriateness of the images.