Staff feared inaccurate machine translation of health messages could harm health and community credibility
Source
Turner (2015). Modeling workflow to design machine translation applications for public health practice. Journal of Biomedical Informatics.
Description #
Several health-department staff expressed concern that inaccurate machine translation of health messages could negatively impact health. Their reservations about adopting MT clustered around five worries: that MT gives literal translations and fails to reflect the nuances of language; concern about the accuracy and fluency of output; fear of damaging relationships with the community through poor-quality translations; concern that MT would not meet policy or program requirements; and a belief that post-translation review of MT might be more work than translating from scratch.
"Several health department staff expressed concern that inaccurate translations of health messages could negatively impact health." (Turner, 2015, p. 142)
"Concerns expressed about the use of MT for translation included: (1) worry that MT may give literal translations and fails to reflect the nuances of language" (Turner, 2015, p. 142)
"Anyway that's what makes everyone nervous about machine translation because we know we'll be raked over the coals every time we get it wrong and we just worry with a machine doing it, the chances of there being an error would be high" (Turner, 2015, p. 143, Table 2)
Methods Context #
What? #ⓘ
The observable: staff-expressed concerns and reservations about using MT for public-health translation, coded as sub-themes of the attitudes-toward-automation theme.
"Several health department staff expressed concern that inaccurate translations of health messages could negatively impact health." (Turner, 2015, p. 142)
How? #ⓘ
Concerns were elicited in the semi-structured interviews and enumerated during thematic coding; representative quotations appear in the themes table (Table 2, sub-theme 9.2).
"Two researchers analyzed the field notes and transcripts using Braun and Clarke's method of thematic analysis [47] to identify broad categories and the initial coding scheme." (Turner, 2015, p. 138)
Who? #ⓘ
Health-department staff among the 34 interviewees at PHSKC and WA DOH who develop and translate health-promotion materials.
"We recruited staff from work units within the departments that frequently performed translations." (Turner, 2015, p. 138)
Other Notes #
This risk-perception finding is distinct from the general "MT quality is poor" appraisal: it specifies why practitioners hesitate to adopt MT (harm to health messaging and community trust), motivating the design requirement that MT errors be tracked and that human review precede public release.