N. Streitz and P. Markopoulos (Eds.): DAPI 2014, LNCS 8530, pp. 262–271, 2014. © Springer International Publishing Switzerland 2014 My Smart TV Agent: Designing Smart TV Persona for Linguistic UX Seyeon Lee, Jiwon Moon, Hokyoung Im, Chung-Kon Shi, and Bong Gwan Jun Graduate School of Culture Technology, KAIST 291 Daehak-ro(373-1 Guseong-dong), Yuseong-gu, Daejeon 305-701, South Korea {birdkite,moon.jiwon,imagine21c,bonggwan,chungkon}@kaist.ac.kr Abstract. Linguistic UX design for Smart TV has been creating much heap as a means of new TV control. Since human voice displays powerful social pres- ence, the issue with defining a Smart TV agent that interacts with users has a big impact in users' satisfaction. The purpose of this study is to analyze the lin- guistic patterns in vocal commands of TV users and to suggest underlying per- sonas of Smart TV agent that appears when users interact with a Smart TV. First, we analyzed most common TV viewing situations and the patterns of us- ers' behavior through a survey. Then, we collected 867 vocal data through a cul- tural probe method in which 10 families, each representing a typical type of TV viewers, by asking them to record what they would like to say to the TV while watching it for about a week. We suggest 6 different type of Smart TV perso- nas, such as expert, assistant, colleague, slave, machine and pet, based on the relationship that the user and TV exhibited. With the collected vocal data, we analyzed the participants' speech pattern and style to examine which type of Smart TV persona was most prevalent. As a result, there were slight difference in types that emerged according different functions of Smart TV and we found that the assistant type appeared most frequently followed by the colleague type. Keywords: Affective communication, Smart TV agent, Linguistic UX, Voice command. 1 Introduction We no longer expect our TV to just change the channel and adjust the volume for us. TVs can now perform a lot more functions with the birth of Smart TV which is capa- ble of bringing the internet to the TV screen [1]. However, it is difficult to encourage users to actively engage with them just because there are more functions to operate since TV is the longest standing Lean Back media in people's mind. That is why there is a steady effort to find the most appropriate way to encourage users to engage in active control of TV utilizing a touch remote control, a motion control, smartphones, motion capture and vocal command etc. The most attractive attribute of vocal command technology is that users can easily pick it up without particular training. However, at the moment, it fails to meet the stan- dard people expect to get from the experience. Of course, the level of communication