"Nobody Speaks that Fast!" An Empirical Study of Speech Rate in Conversational Agents for People with Vision Impairments

Toward the Introduction of Google Assistant in Therapy for Children with Neurodevelopmental Disorders: An Exploratory Study

“…We opted not to directly interview children because questionbased approaches involve critical and self-critical skills [38] that are often lacking in these people [6]. Data analysis took inspiration from a previous study concerning Alexa and another group of people with special needs (i.e., people with visual impairments) [19]. Such as them, we analyzed separately quantitative data (i.e., history logs, part of the forms and the questionnaire) and qualitative data (i.e., the group interview, and part of the forms and the questionnaire) data, but we discussed them together.…”

“…Just once, C6 spontaneously proposed to use the assistant to set the timer during a social game.History logs recorded in total 410 requests to the system. By applying the bottom-up thematic coding approach, we grouped them into 18 categories depending on the intention they convey (see Figure 1), named as usage pattern [19,49]. The most frequent requests were related to "Play animals' cry, or natural rumors" (N = 141), to "Play, or stop music" (N = 92), to set alarms or timers (N = 77), and to tune the volume (N = 28).…”

“…Many studies have been carried out to understand how users adopt an IPA into their daily lives, and social interactions [9,47,55] and to examine the perception of an IPA by users [31,41,50]. Also, much research has been done on IPAs to ofer the potential to support the usage of groups with special needs, such as elderly [48,60,62], children [25], people with vision impairments [19], and military veterans with mild brain injury [57]. However, so far, there is still a limited understanding of how such conversational tools are used or could be used by people with neurodevelopmental disorders (NDD), especially by children.…”

Can we talk? Design Implications for the Questionnaire-Driven Self-Report of Health and Wellbeing via Conversational Agent

“…Similarly, Dubiel et al evaluated the impact of different synthetic voices on participants' perceptions and behavior during a flight-booking task conducted via CA, concluding that although users perceived significant differences between synthetic voices in terms of truthfulness and engagement, this did not translate into significant differences in behavior [18]. Choi et al conducted a 20-day in-home study to investigate visually impaired people's CA use and perceptions of different speech rates [10]. They found that visually impaired users were generally more satisfied with a default human rate of speech, and emphasized speech rate control as an important feature of CA design.…”

“…Research within Human-Computer Interaction (HCI) to date has primarily focused on issues of privacy, usability, user satisfaction and attitudes towards VAs [9,13,15,32]. Recent work has, for example, focused on the impact of CAs' voice characteristics on users' experiences [6,10,18], and the effectiveness of chat-based surveys compared to traditional web-based methods [7,55]. Others have proposed initial guidelines for the design [41,50,54] and evaluation of CAs [26].…”