Unmet Needs and Opportunities for Mobile Translation AI Daniel J. Liebling, Abigail Evans Aaron Donsbach, Michal Lahav Northeastern University Jess Holbrook, Boris Smus, Google Seattle, WA Lindsey Boran Seattle, WA
[email protected] Google dliebling,
[email protected] Seattle, WA donsbach,jessh,smus,
[email protected] ABSTRACT affordances including speech recognition and augmented real- Translation apps and devices are often presented in the con- ity [14]. text of providing assistance while traveling abroad. However, Access to translation, then, appears to be universally available. the spectrum of needs for cross-language communication is However, little public research exists that details people’s trans- much wider. To investigate these needs, we conducted three lation needs and examines how translation apps meet those studies with populations spanning socioeconomic status and needs. For what populations, in which scenarios, and what con- geographic regions: (1) United States-based travelers, (2) mi- tent is translated? How do translation apps succeed or fail at grant workers in India, and (3) immigrant populations in the these tasks? This paper sheds new light on these questions. We United States. We compare frequent travelers’ perception detail the language needs of, and use of translation technology and actual translation needs with those of the two migrant by, three different communities: English-speaking travelers communities. The latter two, with low language proficiency, based in the United States, Hindi-speaking intra-national mi- have the greatest translation needs to navigate their daily lives. grant workers who live in Tamil-speaking Chennai, India, and However, current mobile translation apps do not meet these immigrants with low English proficiency who live in Seattle, needs.