Explore the vast range of titles available.

This study investigates the factors that build resistance and attitude towards AI voice assistants (AIVA). A theoretical model is proposed using the dual-factor framework by integrating status quo bias factors (sunk cost, regret avoidance, inertia, perceived value, switching costs, and perceived threat) and Technology Acceptance Model (TAM; perceived ease of use and perceived usefulness) variables. The study model investigates the relationship between the status quo factors and resistance towards adoption of AIVA, and the relationship between TAM factors and attitudes towards AIVA. A sample of four hundred and twenty was analysed using structural equation modeling to investigate the proposed hypotheses. The results indicate an insignificant relationship between inertia and resistance to AIVA. Perceived value was found to have a negative but significant relationship with resistance to AIVA. Further, the study also found that inertia significantly differs across gender (male/female) and age groupings. The study's framework and results are posited as adding value to the extant literature and practice, directly related to status quo bias theory, dual-factor model and TAM.

Smart homes equipped with ambient intelligence technology constitute a promising direction to enable the growing number of elderly to continue to live in their own home as long as possible. However, this calls for technological solutions that suit their specific needs and capabilities. The Sweet-Home project aims at developing a new user friendly technology for home automation based on voice command. This paper reports a user evaluation assessing the acceptance and fear of this new technology. Eight healthy persons between 71 and 88 years old, 7 relatives (child, grandchild or friend) and 3 professional carers participated in a user evaluation. During about 45 min, the persons were questioned in co-discovery in the Domus smart home alternating between interview and wizard of Oz periods followed by a debriefing. The experience aimed at testing four important aspects of the project: voice command, communication with the outside world, domotics system interrupting a person’s activity, and electronic agenda. Voice interface appeared to have a great potential to ease daily living for elderly and frail persons and would be better accepted than more intrusive solutions. By considering still healthy and independent elderly people in the user evaluation, an interesting finding that came up is their overall acceptance provided the system does not drive them to a lazy lifestyle by taking control of everything. This particular fear must be addressed for the development of smart homes that support daily living by giving them more ability to control rather than putting them away from the daily routine.

The rising population of millennials, coupled with Digital Assistants (DA) and online purchasing trends among consumers have gained increasing attention by global marketers. The study evaluates the influence of DA attributes on the purchasing intention (PUI) of millennials. A combined approach of PLS-SEM, Artificial Neural Network (ANN) and Fuzzy-set Qualitative Comparative Analysis (fsQCA) is used to predict the PUI of 345 millennials. Also, multi-group analysis is employed to uncover the influence of gender on the relationship between PUI and DA attributes. The findings suggest that DA attributes may amplify purchasing intention among millennials, especially through perceived interactivity and anthropomorphism. Further, the moderating role of gender was found significant on the inter-relationship of perceived interactivity and PUI. This research is a pioneer study in the area of artificial intelligence, conversational commerce, DA and AI-powered chatbots. This study will help marketers and practitioners to predict millennial purchasing intentions. An evaluation of this paper may help them to foster immersive and effective engagement through DA.

People with dyslexia have more difficulty than expected in recognizing familiar voices. The ability to recognize people by their voice is an important social behavior. Individuals differ in how they pronounce words, and listeners may take advantage of language-specific knowledge of speech phonology to facilitate recognizing voices. Impaired phonological processing is characteristic of dyslexia and thought to be a basis for difficulty in learning to read. We tested voice-recognition abilities of dyslexic and control listeners for voices speaking listeners’ native language or an unfamiliar language. Individuals with dyslexia exhibited impaired voice-recognition abilities compared with controls only for voices speaking their native language. These results demonstrate the importance of linguistic representations for voice recognition. Humans appear to identify voices by making comparisons between talkers’ pronunciations of words and listeners’ stored abstract representations of the sounds in those words.

Abstract Search modality is becoming increasingly important for Internet platforms and e-commerce businesses. Consumers can perform product searches on the Internet by typing their search queries (typed search modality) or by speaking them (voice search modality). Given the variation and the managerial ease of selecting different search modalities to adopt, we investigate the consequences of search modalities on consumers’ mindsets and purchase intentions. Six studies, including an Implicit Association Test and an incentive-compatible field experiment, show that typed search modality (vs. voice search modality) led to higher purchase intentions and behavior. This results from learned responses where typing is nonconsciously associated with taking action, and vocalization is nonconsciously associated with information gathering and deliberation. Thus, consumers performing a typed search are more likely to be in an action-oriented mindset, whereas consumers performing a voice search are more likely to be in a deliberative mindset. Our research carries implications for digital technologies.

Anthropomorphic voice assistants (e.g., Amazon Alexa) enable users to use natural-language voice commands to control “smart” objects and access the internet for information, shopping, and entertainment. Most manufacturers of voice assistants allow other firms to develop software (i.e., voice assistant functions, VAFs) related to their products and services that add new capabilities to voice assistants. To measure the value of different types of capabilities of VAFs, we empirically study the impact of announcements of VAFs on firm value. We show that informational capabilities and VAFs announced by product firms have a positive moderating effect on firm value. On the other hand, object-control capabilities have no moderating impact on firm value, while transactional capabilities have a negative impact. Theoretical and managerial implications are discussed. Additionally, necessary avenues for future research within the voice assistant domain are proposed.

Language recognition systems based on bottleneck features have recently become the state-of-the-art in this research field, showing its success in the last Language Recognition Evaluation (LRE 2015) organized by NIST (U.S. National Institute of Standards and Technology). This type of system is based on a deep neural network (DNN) trained to discriminate between phonetic units, i.e. trained for the task of automatic speech recognition (ASR). This DNN aims to compress information in one of its layers, known as bottleneck (BN) layer, which is used to obtain a new frame representation of the audio signal. This representation has been proven to be useful for the task of language identification (LID). Thus, bottleneck features are used as input to the language recognition system, instead of a classical parameterization of the signal based on cepstral feature vectors such as MFCCs (Mel Frequency Cepstral Coefficients). Despite the success of this approach in language recognition, there is a lack of studies analyzing in a systematic way how the topology of the DNN influences the performance of bottleneck feature-based language recognition systems. In this work, we try to fill-in this gap, analyzing language recognition results with different topologies for the DNN used to extract the bottleneck features, comparing them and against a reference system based on a more classical cepstral representation of the input signal with a total variability model. This way, we obtain useful knowledge about how the DNN configuration influences bottleneck feature-based language recognition systems performance.

Unlike most other computing devices that are known to isolate their users, Smart Voice Assistant Speaker (SVAS) appears to improve the perception of social cohesion (i.e., Group Harmony) among its co-users. We hypothesize that the social cues emanated from the continued, and habituated, use of SVAS develop the “illusion of intimacy” which, in turn, ripples through the entire group, and help fulfill the need for social integration. The data collected from 218 families support this hypothesis. We argue that just as a puppy dog contributes to a happy family, so does the SVAS contributes to the social dynamics by making the users unconsciously fulfill their psychological needs and by increasing actual conversations among its users. Incidentally, the study compared the relative influence of three factors (the beta weight of “Hedonic Motivation” being the highest, followed by “Compatibility,” and then “Perceived Security”) that, as a whole, explain over 60% of the variance in Satisfaction of post-adoption SVAS use.

Background To undertake a systematic review of existing literature relating to speech recognition technology and its application within health care. Methods A systematic review of existing literature from 2000 was undertaken. Inclusion criteria were: all papers that referred to speech recognition (SR) in health care settings, used by health professionals (allied health, medicine, nursing, technical or support staff), with an evaluation or patient or staff outcomes. Experimental and non-experimental designs were considered. Six databases (Ebscohost including CINAHL, EMBASE, MEDLINE including the Cochrane Database of Systematic Reviews, OVID Technologies, PreMED-LINE, PsycINFO) were searched by a qualified health librarian trained in systematic review searches initially capturing 1,730 references. Fourteen studies met the inclusion criteria and were retained. Results The heterogeneity of the studies made comparative analysis and synthesis of the data challenging resulting in a narrative presentation of the results. SR, although not as accurate as human transcription, does deliver reduced turnaround times for reporting and cost-effective reporting, although equivocal evidence of improved workflow processes. Conclusions SR systems have substantial benefits and should be considered in light of the cost and selection of the SR system, training requirements, length of the transcription task, potential use of macros and templates, the presence of accented voices or experienced and in-experienced typists, and workflow patterns.

Most totally blind people have non-24-hour sleep–wake disorder (non-24), a rare circadian rhythm disorder caused by an inability of light to reset their circadian pacemaker. In two consecutive placebo-controlled trials (SET and RESET), we assessed safety and efficacy (in terms of circadian entrainment and maintenance) of once-daily tasimelteon, a novel dual-melatonin receptor agonist. We undertook the placebo-controlled, randomised, double-masked trials in 27 US and six German clinical research centres and sleep centres. We screened totally blind adults (18–75 years of age), who were eligible for the randomisation phase of SET if they had a non-24-hour circadian period (τ) of 24·25 h or longer (95% CI greater than 24·0 and up to 24·9 h), as calculated from measurements of urinary 6-sulphatoxymelatonin rhythms. For SET, we used block randomisation to assign patients (1:1) to receive tasimelteon (20 mg) or placebo every 24 h at a fixed clock time 1 h before target bedtime for 26 weeks. Patients who entered the open-label group receiving tasimelteon in SET or who did not meet the SET inclusion criteria but did meet the RESET inclusion criteria were screened for RESET. A subset of the patients who entered the open-label group before the RESET study and who had eligible τ values were screened for RESET after completing the open-label treatment. In RESET, we withdrew tasimelteon in a randomised manner (1:1) in patients who responded (ie, entrained) after a tasimelteon run-in period. Entrainment was defined as having τ of 24·1 h or less and a 95% CI that included 24·0 h. In SET, the primary endpoint was the proportion of entrained patients, assessed in the intention-to-treat population. The planned step-down primary endpoint assessed the proportion of patients who had a clinical response (entrainment at month 1 or month 7 plus clinical improvement, measured by the Non-24 Clinical Response Scale). In RESET, the primary endpoint was the proportion of non-entrained patients, assessed in the intention-to-treat population. Safety assessments included adverse events and clinical laboratory measures, assessed in all treated patients. These trials are registered with ClinicalTrials.gov, numbers NCT01163032 and NCT01430754. Between Aug 25, 2010, and July 5, 2012, we screened 391 totally blind patients for SET, of whom 84 (22%) were assigned to receive tasimelteon (n=42) or placebo (n=42). Two patients in the tasimelteon group and four in the placebo group discontinued the study before τ was measured, due to adverse events, withdrawal of consent, and a protocol deviation. Circadian entrainment occurred in eight (20%) of 40 patients in the tasimelteon group compared with one (3%) of 38 patients in the placebo group at month 1 (difference 17%, 95% CI 3·2–31·6; p=0·0171). Nine (24%) of 38 patients showed a clinical response, compared with none of 34 in the placebo group (difference 24%, 95% CI 8·4–39·0; p=0·0028). Between Sept 15, 2011, and Oct 4, 2012, we screened 58 patients for eligibility in RESET, 48 (83%) of whom had τ assessed and entered the open-label tasimelteon run-in phase. 24 (50%) patients entrained, and 20 (34%) were enrolled in the randomisation phase. Two (20%) of ten patients who were withdrawn to placebo remained entrained compared with nine (90%) of ten who continued to receive tasimelteon (difference 70%, 95% CI 26·4–100·0; p=0·0026). No deaths were reported in either study, and discontinuation rates due to adverse events were comparable between the tasimelteon (3 [6%] of 52 patients) and placebo (2 [4%] of 52 patients) treatment courses. The most common side-effects associated with tasimelteon in SET were headache (7 [17%] of 42 patients given tasimelteon vs 3 [7%] of 42 patients given placebo), elevated liver enzymes (4 [10%] vs 2 [5%]), nightmares or abnormal dreams (4 [10%] vs none), upper respiratory tract infection (3 [7%] vs none], and urinary tract infections (3 [7%] vs 1 [2%]). Once-daily tasimelteon can entrain totally blind people with non-24; however, continued tasimelteon treatment is necessary to maintain these improvements. Vanda Pharmaceuticals.