ChatGPT翻译K10、PHQ-9等心理健康量表为中文结果不一致
ChatGPT's translation of mental health questionnaires into Mandarin reveals inconsistent results
坎特伯雷大学(UC)Dr. Wei Teng研究发现,ChatGPT将K10、PHQ-9等心理健康筛查工具翻译为普通话时存在一致性问题。同一提示词在不同账户中产生的译文在语气和自然程度上存在差异。研究者指出,心理健康评估的一致性至关重要,AI在此领域的应用仍需临床监督与文化敏感性。
As more Asian New Zealanders report experiencing mental health challenges, a Te Whare Wānanga o Waitaha | University of Canterbury (UC) researcher is asking a critical question: Can artificial intelligence (AI) help bridge language barriers when human translation resources are limited?
Research led by UC scholar of translation studies Dr. Wei Teng examined how effectively ChatGPT can translate two widely used mental health screening instruments—the Kessler Psychological Distress Scale (K10) and the Patient Health Questionnaire-9 (PHQ-9)—into Mandarin Chinese for use in Aotearoa New Zealand.
"These scales influence whether someone is identified as experiencing distress or depression, so if there's something incorrect in the translation, that directly affects clinical decisions," Teng says.
Teng, a passionate accessibility advocate, began the project after conversations with Josiah Jennings from Christchurch-based Asian Community Transformation Trust and registered clinical psychologist Dr. Amy Wang. During these sessions, community leaders raised concerns about language barriers preventing Mandarin-speaking migrants from accessing appropriate mental health support.
While Mandarin versions of these tools already exist, Teng says some translations feel overly formal or unnatural. "If the translation of questions to Mandarin sounds robotic and loses the tone of the source text, patients may not respond openly."
To test whether AI could produce more natural, empathetic translations, Teng designed structured prompts for ChatGPT, assigning it a professional role and specifying tone requirements such as "caring" and "clinically appropriate." One carefully engineered prompt produced noticeably stronger results, suggesting AI could enhance translation quality when guided more precisely.
However, the study revealed a critical flaw: inconsistency.
When colleagues ran the same prompt through their own free ChatGPT accounts, the outputs differed in tone and naturalness. Because large language models generate responses based on account users' previous interactions, results can vary between users, even when using identical instructions.
"In mental health, consistency matters," Teng says. "A screening tool must function the same way each time. Variability raises questions about reliability and reproducibility."
Now working with AI researchers, Teng is exploring whether more deterministic prompting or customized models could improve consistency. He stresses that while AI shows promise, clinical oversight and interdisciplinary collaboration are essential.
"There may be room for AI in this space," he says. "But when people's well-being is involved, trust and cultural sensitivity must come first."
Who's behind this story?
Lisa Lock
BA art history, MA material culture. Former museum editor, paramedic, and transplant coordinator. Editing for Science X since 2021. Full profile →
Andrew Zinin
Master's in physics with research experience. Long-time science news enthusiast. Plays key role in Science X's editorial success. Full profile →
Citation: ChatGPT's translation of mental health questionnaires into Mandarin reveals inconsistent results (2026, November 19) retrieved 19 November 2026 from https://medicalxpress.com/news/2026-11-chatgpt-mental-health-questionnaires-mandarin.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
来源:Medical Xpress · medicalxpress.com