User experience quality (UEQ) plays a critical role in multilingual intelligent voice assistant (VA), where real-time feedback directly affect interaction experience. However, current UEQ evaluation methods primarily depend on surveys, posing limitations for dynamic and user-centered adaptation. We propose a non-intrusive real-time system, UEQManager, for automated UEQ recognition and management. First, UEQManager incorporate interpretable deep learning models to predict UEQ in seven subdimensions. Second, leveraging LLM and expert focus group, we design adaptive interaction interface based on recognized UEQ states. Third, we implement integration system capable of real-time, non-intrusive UEQ recognition and management. We validate the effectiveness of UEQManager through a user testing experiment. Results reveal that UEQManager significantly outperformed baseline, yielding an average UEQ improvement of 27.29% over baseline and demonstrating statistical significance across all subdimensions. This work contributes a proof of concept HCI system that translates webcam gaze cues into adaptive design decisions for multilingual VAs and illustrates how designers can couple interpretable sensing with proactive interaction design.

The comprehensibility of speech has been extensively studied in academia. Modern intelligent voice assistants are evaluated on the basis of metrics such as comprehensibility, sound quality, and accuracy. However, few studies have examined the impact of emergencies on the comprehensibility of intelligent voice assistants. Additionally, how adjustments to text length and other factors related to intelligent voice assistants can promote optimal comprehensibility, trustworthiness, and compliance during emergencies remains unclear. This study aims to investigate the effects of text length and speech rate on the comprehensibility of intelligent voice assistants during emergencies. An empirical study was conducted with 60 participants (30 males and 30 females). The experiment included 3 text lengths (short: 15–20 words; medium: 30–35 words; long: 45–50 words), 7 speech rates (L1–L7: 250–430 words per minute, with 30 words per minute intervals), and 2 types of emergencies (earthquakes and typhoons). Seven measures were used to evaluate the performance of intelligent voice assistants during emergencies: comprehensibility (self-reported accuracy), intelligibility (accuracy, scored post-transcription via word-by-word comparison and four machine learning methods), trustworthiness, compliance, the perceived sound quality of the intelligent voice assistant, the perceived accuracy of the intelligent voice assistant, and skin conductance. The results indicated that text length, speech rate, and the type of emergency significantly affected the seven measures. Specifically, short texts led to better performance than did long texts across all the measures; the effect of speech rate exhibited an M-shaped curve, with peaks at L2 (280 words per minute) and a “gray zone” at L4 (340 words per minute), where performance was lower than that at L3 and L5. The levels of comprehensibility and intelligibility were higher during earthquakes than during typhoons, but trustworthiness, compliance, and skin conductance were lower during earthquakes. On the basis of these findings, three design recommendations for intelligent voice assistants during emergencies are proposed.

This study experimented with mobile devices to investigate the effects of tones and time pressure on the voice input experience for Mandarin Chinese. Voice input for Mandarin Chinese is gaining popularity, and its implementation on mobile devices has recently developed very rapidly. Previous studies have focused mainly on technical considerations for designing a voice input system and have not investigated the user experience of voice input. Two experiments were conducted to test single-character and two-character phrases. Four measurements, the number of errors, user satisfaction, mental workload, and skin conductance, were gathered from sixty-four participants for each experiment. The results indicated that Mandarin tones had a significant influence on the number of errors and user satisfaction, and time pressure significantly influenced all measurements. Specifically, the falling-rising tone produced the highest number of errors, and the rising tone led to the lowest satisfaction. A high level of time pressure resulted in low satisfaction, a high mental workload, and high skin conductance. The results indicated that (a) Mandarin tones significantly affected error rates and user satisfaction, and (b) time pressure had a significant effect on user experience. The findings have design implications for voice input systems in terms of user experience.

Information sharing is of great importance in the financial market. Most current electronic transaction systems do not encourage investors to freely share information such as risk assessment records. Blockchain, a decentralised technology with high security features, is expected to improve the current systems. However, little is known about whether blockchain-based electronic transaction systems contribute to investors’ willingness to share personal financial information. This study adopts a vignette-based experiment to enrich previous studies by explaining the relationship between blockchain and investors’ behaviour. This study introduced investor empowerment, trust belief, risk belief, and perceived uncertainty as mediating variables and considered how these mediating variables interact with each other. Independent sample t-test and structural equation modelling were used to examine the causal relationship between independent and dependent variables, and the role of mediating effect. For text analysis of interview dialogues, keyword extraction and sentiment analysis were used to understand the participants’ focus and sentiment attitude toward electronic transaction systems based on blockchain and non-blockchain. The results show that blockchain can build trust, reduce risk belief and perceived uncertainty by empowering investors, thus increasing investors’ information sharing intention and investment intention. Furthermore, investors have significantly more positive emotional attitudes toward blockchain-based systems.

Voice interaction is a user-friendly interaction method for Mandarin Chinese. Voice perception has been widely implemented on smart devices for Mandarin voice interaction. However, previous studies gave little attention to how language features, such as the text type and text length affect the voice perception experience. Moreover, how the device type influences Mandarin voice perception is unclear, as the device type has been shown to have a significant effect on other interaction methods, such as interactions with keyboard entry and screen displays. Thus, the aim of this study was to investigate the effects of the text type, text length, and device type on the voice perception experience. This study invited seventy-two participants (36 females and 36 males) for an empirical experiment. Three text types, i.e., classical Chinese text, professional text, and daily text, three text lengths, namely, short (i.e., 10 characters), medium (i.e., 20 characters), and long (i.e., 30 characters), and two device types, i.e., mobile phones and smartwatches, were considered. Four measurements were used to evaluate the voice perception experience, namely, performance (i.e., self-reported performance and correct ratio), the satisfaction rating, mental workload, and skin conductance. The results indicate that the text type, text length, and device type had significant effects on the four measurements. Specifically, the voice perception experience was best for daily text and worst for classical Chinese text, shorter text led to a better voice perception experience, and users of smartwatches had a better voice perception experience than mobile phones. Furthermore, the design implications for improving the voice perception experience that consider the text type, text length, and device type were discussed.

Handwriting is widely used by Chinese older adults for text input on touch screens, but their handwriting performance is relatively low in comparison to younger adults. Therefore, this study investigated how the size of the input box (15%, 30%, and 45% of the display), input method (fingers and stylus), display size (4.7 inches and 7.9 inches), and age (younger adults and older adults) influence Chinese handwriting performance and user experience. An experiment was conducted among 72 participants. A system was developed to automatically record the time of handwriting, the accuracy of handwriting, the number of protruding strokes, the number of rewritings. Besides, participants’ satisfaction, preferences, and mental workload were measured through questionnaires. The results revealed significant effects of the size of the input box, age, input method, and display size on the performance indexes and subjective ratings. Accordingly, 12 design implications for improving the Chinese handwriting system were identified.

There is a growing tendency for users to expect conversational agents (CAs) to recognise social cues and follow interpersonal communication principles to enhance their subjective evaluation. Therefore, this paper studies how personal pronouns should be used by CAs in response to users. We conducted a 3 (CAs’ personal pronoun) × 3 (users’ personal pronoun) × 2 (participants’gender) mixed design. this study used mixed methods based on an experimental design, including ratings, forced choices and interviews, for mutual confirmation. The findings indicate that first, users prefer that CAs use second-person pronouns. Second, there is also turn-taking and convergence tendency between users and CAs in personal pronoun use. Third, there are gender differences in personal pronoun preferences and relationship positions toward CAs. These results can inform personalised voice interaction and humanlike design and help build closer relationships between users and CAs in future human–computer interactions.