Yao Zhang.

I translated field feedback into design requirements for satellite communications. After integration, we ran ground and flight usability tests at Cijulang Airport in Indonesia. I surveyed 30 onsite operators and engineers after functional validation to gather feedback for the next iteration.

I interviewed Thai and Indonesian operators and collaborated with Chinese engineers. I designed structured prompts for Chinese-to-English abbreviations using aviation terminology and interface length constraints. Three domain experts reviewed the outputs, and 25 operators from Thailand, Indonesia, and the UAE took part in simulation tasks to identify ambiguous wording and moments of hesitation.

I recruited 24 participants, eight native speakers each of Mandarin, Korean, and Japanese. Each participant began ten tasks in ChatGPT Voice with a scripted English request, then chose how to respond if the system misunderstood them. I recorded the interactions, asked about trust at three points, and used video playback in semi-structured interviews to explore participants’ choices.

The analysis combines interaction coding, descriptive statistics, and reflexive thematic analysis. It compares observed actions, such as rephrasing a request or switching to typing, with participants’ explanations of whether those actions felt like compromises. Task completion and self-reported trust are analyzed separately.