By Amira Asfar
Interviews and Real-World Simulations
Researchers at ZHAW Zurich University of Applied Sciences tested whether machine interpreting apps like Microsoft Translator can support everyday communication between refugees and professionals.
As part of the study, four semi-guided interviews were conducted with non-linguist professionals (a job coach, migration office manager, school social worker, and midwife) who work daily with refugees. The interviews explored technology use, the advantages and drawbacks of human interpreters, and the ideal features for translation apps.
In addition, the researchers carried out four simulated interactions between Ukrainian refugees (with limited German proficiency) and Swiss professionals. These simulations recreated realistic scenarios, such as: informing a parent about a school trip, assessing employability for job entry, triaging a bike accident injury and advising on German language course enrolment.
Participants used a tablet or phone with Microsoft Translator in a face-to-face setup. Each conversation lasted around 20 minutes and was recorded for analysis of verbal, nonverbal, and technology-handling behaviours. Afterward, post-task interviews captured participants’ feedback and reflections.
Key Findings
- Availability vs. Trust: Human interpreters are valued for cultural insight but often unavailable or expensive. Machine interpreting offers independence but raises data privacy concerns.
- User Experience: Both refugees and professionals rated machine interpreting positively. Refugees felt more empowered to participate equally, breaking passive roles common with human interpreters.
- Accuracy: Translation errors occurred (e.g., gender bias, misread abbreviations), but users adapted quickly and rarely saw them as barriers.
- Conversation Flow: Apps disrupted natural pacing, for ex. cutting off slow speakers or forcing shorter turns, but participants developed coping strategies.
- Nonverbal and Cognitive Aspects: Eye contact dropped, but gestures and nods replaced it as new communication cues. Professionals experienced mental fatigue; refugees gained agency.
- Handling Aspects: Table placement and device handling influenced interaction comfort. Hands-free or wearable setups could help.
What this means for your work
If you’re in refugee services, this study is a nudge to experiment with machine interpreting for low-stakes chats (e.g., school updates, not asylum hearings). It could cut costs, boost independence, and scale support amid interpreter shortages. But don’t go all-in without guidelines: equip staff with training on technological nuances, prioritize tools that ensure data privacy, and integrate human support for high-stakes situations.
The researchers call for more studies on emerging practices, like tech-adapted turn-taking, and better designs (e.g., reader-friendly text displays). With AI like Large Language Models advancing, these tools could get safer and smarter soon, potentially transforming SDGs from aspiration to action.
Curious? Check the full paper (DOI: https://doi.org/10.14201/0AQ0373337347) for references and additional insights.