Voice Agent Research Articles

Plant Robot for At-Home Behavioral Activation Therapy Reminders to Young Adults with Depression

Adolescents with depression who participate in behavioral activation therapy may find it hard to be motivated to perform tasks at home that their therapists recommend. We describe the initial design and usability evaluation of a home device (“PlantBot”) that could be used to remind young adults with depression at home of their behavioral activation therapy-related tasks. The prototype features electronics in a two-layer base, with a fake plant on top and supported using the Amazon Echo voice agent. We use an online panel study to evaluate the usability of our system with youth with past depression (N= 30). Initial findings highlight the device’s usability, potential benefit, and attractiveness of the plant component, as well as multiple improvements to be made.

Voice Technologies to Support Aging in Place: Opportunities and Challenges

Technology to support aging in place has been a topic of interest. Research indicates that older adults are increasingly using commercially available voice assistants in smart speakers. These devices enable non-visual interaction that does not require extensive expertise with traditional mobile or desktop computers, thus offering new possibilities of access to digital technology. We conducted two different studies with individuals aged 65 years old or above—a three week smart speaker deployment study with individuals who did not use computing devices regularly and a workshop on customizing internet of things technology with tech savvy individuals. Our findings indicate specific ways that these voice technologies might support aging in place, including ease of use and due to their not being identified with aging-specific technologies. We observed that participants consistently used their voice agent for finding online information, particularly health-related, emphasizing the need to revisit concerns about credibility of information with this new interaction medium. And, although features to support memory (e.g., setting timers, reminders) were initially perceived as useful, the actual usage was unexpectedly low due to reliability concerns. Our work provides a basis to understand older adults’ perceptions and usage of current voice technologies. We also identify opportunities for customizing voice technologies to better support aging in place.

Female Voice Agents in Fully Autonomous Vehicles Are Not Only More Likeable and Comfortable, But Also More Competent

Driving agents can provide an effective solution to improve drivers’ trust in and to manage interactions with autonomous vehicles. Research has focused on voice-agents, while few have explored robot-agents or the comparison between the two. The present study tested two variables - voice gender and agent embodiment, using conversational scripts. Twenty participants experienced autonomous driving using the simulator for four agent conditions and filled out subjective questionnaires for their perception of each agent. Results showed that the participants perceived the voice only female agent as more likeable, more comfortable, and more competent than other conditions. Their final preference ranking also favored this agent over the others. Interestingly, eye-tracking data showed that embodied agents did not add more visual distractions than the voice only agents. The results are discussed with the traditional gender stereotype, uncanny valley, and participants’ gender. This study can contribute to the design of in-vehicle agents in the autonomous vehicles and future studies are planned to further identify the underlying mechanisms of user perception on different agents.

스마트 홈 맥락에서 음성 에이전트에 대한 사용자 니즈 연구

스마트 홈 맥락에서 음성 에이전트에 대한 사용자 니즈 연구

카셰어링 서비스 환경에서 차량 내 음성 에이전트의 관계적인 개인화 수준과 재사용에 따른 사용자 경험 효과 연구

공유경제의 흐름이 자동차 산업에도 영향을 미치면서 카셰어링 서비스가 크게 성장하고 있다. 차량을 일시적으로 사용하는 카셰어링 서비스에 대해 긍정적인 사용자경험(UX)을 제공하기 위해서는 개인화된 인터페이스에 대한 연구가 필요하다. 본 연구는 대인관계 상호작용 원리를 카셰어링 서비스 차량 내 음성 에이전트에 적용하여, 관계적인 개인화가 사용자의 장소애착, 심리적 소유감, 신뢰와 지속 사용 의도에 영향을 주는지 살펴보았다. 또한 카셰어링 서비스는 재사용 시마다 다른 차량을 이용하게 되는데, 차량 내 음성 에이전트의 관계적인 개인화가 재사용 시 영향을 미치는지 확인하였다. 분석 결과, 관계적인 개인화가 형성된 대상과의 동행은 처음 방문하는 낯선 공간이라 할지라도 사용자의 정서에 긍정적인 영향을 준다는 가능성을 확인하였다.

Does the effect of enthusiasm in a pedagogical Agent's voice depend on mental load in the Learner's working memory?

The emotional design hypothesis states that features in digital learning materials can elicit emotional reactions. Voice-related cues, such as enthusiastic accentuations are getting more into the focus of research. These emotion-related cues have to be additionally processed to the relevant instructions while learning, so that the overall amount of to-be-processed information might be a moderator of the effect of emotional cues. One hundred eighteen participants were assigned to one cell of a two (enthusiasm of a pedagogical agent: enthusiastic vs. neutral) × two (mental load in the working memory of the learner: high vs. low) factorial between-subjects design. Regarding the multiple-choice learning test, results revealed that learners with a neutral agent voice performed better in a high load condition, while learners exposed to an enthusiastic voice reached higher learning scores in a low load condition. However, this moderating effect of mental load could not be shown for open answer questions.

Exploring Smart Agents for the Interaction with Multimodal Mediated Environments

After conversational agents have been made available to the broader public, we speculate that applying them as a mediator for adaptive environments reduces control complexity and increases user experience by providing a more natural interaction. We implemented and tested four agents, each of them differing in their system intelligence and input modality, as personal assistants for Mediated Atmospheres, an adaptive smart office prototype. They were evaluated in a user study ( N = 33 ) to collect subjective and objective measures. Results showed that a smartphone application was the most favorable system, followed by conversational text and voice agents that were perceived as being more engaging and intelligent than a non-conversational voice agent. Significant differences were observed between native and non-native speakers in both subjective and objective measures. Our findings reveal the potential of conversational agents for the interaction with adaptive environments to reduce work and information overload.

Effects of personalization and social role in voice shopping: An experimental study on product recommendation by a conversational voice agent

This study examines the persuasion mechanism in product recommendations made by a voice-based conversational agent and explores whether the personalized content reflecting the customer's preferences and the agent's social role of a friend, rather than a secretary, generate a more positive attitude toward the product in the context of voice shopping. With the framework of dual modes of information-processing models, we hypothesized that the personalization of messages would be a central route with a greater impact on attitude for products with high involvement. By contrast, the social role of the conversational agent was expected to represent a peripheral route with a greater impact on products with low involvement. An experimental study was designed to test the effects of personalized content that reflected individual preferences for product attributes and a friend role of a voice agent with high and low product involvement. The results showed main effects of both personalization and the social role on building attitudes toward the product. Although no interaction effect for personalization and involvement was found, there was a significant interaction effect for the social role and involvement. This study contributes to persuasion theory by extending it to the interaction with a conversational agent. For practitioners, the study provides insights into the importance of the personalized content of recommendations and the need for consideration of an alternative social role in the design of voice shopping through a conversational agent.

Merging and synchronizing corporate and personal voice agents: Comparison of voice agents acting as a secretary and a housekeeper

People are currently facing a one-to-multiple mode of interaction with voice agents (VAs), which are widely used in their personal and working lives separately owing to the rapid development of artificial intelligence. To explore potential trends in the use of VAs, we designed a merged VA and two either familiar (interconnected) or strange (completely independent) VAs, of which one acted as a secretary for work-related issues and the other acted as a housekeeper for personal issues. Seventy-two people participated in a within-group experiment to experience simulated interactions and complete tasks using the designed VAs in a random order. Objective and subjective evaluation data were collected to analyse user experience. The results show that the merged VA performs best and that participants also preferred it over the others based on the collected data on mental workload, performance, satisfaction, and perceived usefulness. The familiar VAs were the second best-performing VAs. Though there were no significant differences in the trust in the three designed VAs, participants held positive and cautious attitudes towards VAs in general. Additionally, this study highlights the importance of efficiency, the value of entertainment, and the distinction between work and personal life with respect to the personal and corporate functions of VAs.

Survey on Psychological Effects of Appreciative Communication via Voice Agent

Survey on Psychological Effects of Appreciative Communication via Voice Agent

"Phantom Friend" or "Just a Box with Information"

As voice-based conversational agents such as Amazon Alexa and Google Assistant move into our homes, researchers have studied the corresponding privacy implications, embeddedness in these complex social environments, and use by specific user groups. Yet it is unknown how users categorize these devices: are they thought of as just another object, like a toaster? As a social companion? Though past work hints to human-like attributes that are ported onto these devices, the anthropomorphization of voice assistants has not been studied in depth. Through a study deploying Amazon Echo Dot Devices in the homes of older adults, we provide a preliminary assessment of how individuals 1) perceive having social interactions with the voice agent, and 2) ontologically categorize the voice assistants. Our discussion contributes to an understanding of how well-developed theories of anthropomorphism apply to voice assistants, such as how the socioemotional context of the user (e.g., loneliness) drives increased anthropomorphism. We conclude with recommendations for designing voice assistants with the ontological category in mind, as well as implications for the design of technologies for social companionship for older adults.

The Voice Makes the Car: Enhancing Autonomous Vehicle Perceptions and Adoption Intention through Voice Agent Gender and Style

The present research explores how autonomous vehicle voice agent (AVVA) design influences autonomous vehicle passenger (AVP) intentions to adopt autonomous vehicles. An online experiment (N = 158) examined the role of gender stereotypes in response to an AVVA with respect to the technology acceptance model. The findings indicate that characteristics of the AVVA that are more consistent with the stereotypical expectation of the social role (informative male AVVA and social female AVVA) foster greater perceived ease of use (PEU) and perceived usefulness (PU) than inconsistent conditions (social male AVVA and informative female AVVA). The study offers theoretical implications regarding the technology acceptance model in the context of autonomous technologies as well as practical implications for the design of autonomous vehicle voice agents.

Designing for Trust: Understanding the Role of Agent Gender and Location on User Perceptions of Trust in Home Automation

Although trust can have a positively mediating effect on information technology adoption and usage, the concept has not been extensively investigated in the home automation field. Therefore, this work is aimed at exploring the role of agent location and the gender of the agent's voice on users' perception of trust toward automation through two experimental studies (N = 8 and N = 20) and a web-based smart lock simulation. Explicit trust behavior was captured using directly observable behaviors and decisions, while implicit trust behavior was captured using detailed click-level user behaviors with the smart lock simulation as a proxy for reaction time. The results show that users displayed more explicit trusting behavior toward the system when it displayed design characteristics that were stereotype congruent (female-home and male-office) compared to stereotype incongruent systems (male-home and female-office). These results show that users carry over the social expectations and roles encountered in human-to-human relationships to interactions with simulated automated agents. These findings empirically demonstrate the influence of design characteristics on the formation of trust relationships between users and automated devices and provide a foundation for future research geared at critically examining our evolving relationship with technology.

“Driver Take Over”: A Preliminary Exploration of Driver Trust and Performance in Autonomous Vehicles

Automated vehicles are becoming more prominent in research and development. These automated vehicles introduce issues that have been seen in other autonomous systems such as decreases in situation awareness, complacency, and trust. Previous literature has looked at the effects of alerts and voice agents on driving performance. This preliminary study compares different in-car alerts (no alert, sound alert, task irrelevant voice alert, and task relevant voice alert) on trust and the driver’s ability to get back in-the-loop when the automation has failed. Participants were asked to monitor a simulated automated vehicle as it drove down a straight two-lane road. The main statistical results of our study show no difference in trust between the four different conditions; however, more participants avoided collision with a leading car in the task relevant voice condition in comparison to the three other conditions. These preliminary findings have important implications for the design of automated vehicles.

A Multimodal Approach to Assessing User Experiences with Agent Helpers

The study of agent helpers using linguistic strategies such as vague language and politeness has often come across obstacles. One of these is the quality of the agent's voice and its lack of appropriate fit for using these strategies. The first approach of this article compares human vs. synthesised voices in agents using vague language. This approach analyses the 60,000-word text corpus of participant interviews to investigate the differences of user attitudes towards the agents, their voices and their use of vague language. It discovers that while the acceptance of vague language is still met with resistance in agent instructors, using a human voice yields more positive results than the synthesised alternatives. The second approach in this article discusses the development of a novel multimodal corpus of video and text data to create multiple analyses of human-agent interaction in agent-instructed assembly tasks. The second approach analyses user spontaneous facial actions and gestures during their interaction in the tasks. It found that agents are able to elicit these facial actions and gestures and posits that further analysis of this nonverbal feedback may help to create a more adaptive agent. Finally, the approaches used in this article suggest these can contribute to furthering the understanding of what it means to interact with software agents.

An Agent-Based Intervention to Assist Drivers Under Stereotype Threat: Effects of In-Vehicle Agents' Attributional Error Feedback.

For members of a group negatively stereotyped in a domain, making mistakes can aggravate the influence of stereotype threat because negative stereotypes often blame target individuals and attribute the outcome to their lack of ability. Virtual agents offering real-time error feedback may influence performance under stereotype threat by shaping the performers' attributional perception of errors they commit. We explored this possibility with female drivers, considering the prevalence of the "women-are-bad-drivers" stereotype. Specifically, we investigated how in-vehicle voice agents offering error feedback based on responsibility attribution (internal vs. external) and outcome attribution (ability vs. effort) influence female drivers' performance under stereotype threat. In addressing this question, we conducted an experiment in a virtual driving simulation environment that provided moment-to-moment error feedback messages. Participants performed a challenging driving task and made mistakes preprogrammed to occur. Results showed that the agent's error feedback with outcome attribution moderated the stereotype threat effect on driving performance. Participants under stereotype threat had a smaller number of collisions when the errors were attributed to effort than to ability. In addition, outcome attribution feedback moderated the effect of responsibility attribution on driving performance. Implications of these findings are discussed.

Can “The Voices in the Car” Persuade Drivers to Go Green?: Effects of Benefit Appeals from In-Vehicle Voice Agents and the Role of Drivers' Affective States on Eco-Driving

The present research investigated the possibility of using an in-vehicle voice agent to promote eco-driving. Considering that both types of benefit appeals--egoistic (emphasizing benefits to the self) and altruistic (emphasizing benefits to others)--could be employed to promote eco-driving behavior, we explored the effects of benefit appeals delivered by an in-vehicle voice agent on driving performance. In particular, we tested whether and how the valence (positive vs. negative) of drivers' affective states moderates the effects, drawing on the functionalist affect-cognition framework, which has theorized that positive affect leads people to focus more on self-interest, whereas negative affect leads people to become more sensitive to social norms. An experiment was conducted in which participants, after undergoing affect (happy vs. sad) elicitation, received messages (egoistic vs. altruistic) promoting eco-driving from an in-vehicle voice agent while performing a simulated driving task. Results were partially consistent with the functionalist affect-cognition framework. Happy participants performed better on eco-driving when they were exposed to egoistic appeals than to altruistic appeals. On the other hand, the driving performance data from sad participants did not yield a significant difference between the egoistic condition and the altruistic condition. Participants' driving performance data further revealed that the joint effects of benefit appeals and affective states on safe driving performance mirrored the joint effects on eco-driving performance, confirming a close relationship between the two driving behaviors. Theoretical and practical implications for the use of in-vehicle voice agents and benefit appeals in promoting eco-driving and safe driving are discussed.

The performance of a newly developed mobile hands-free context-aware communications system in a hospital environment

This paper characterizes the effect of hospital noise conditions on a mobile communications system developed for a research trial in a Johns Hopkins Hospital nursing unit. The system provides an alternative to telephones and overhead paging in hopes of reducing the amount of noise on the hospital unit. Benefits of this system over current similar solutions include a completely hands free approach using a computerized voice agent and a design that addresses privacy concerns to comply with HIPAA requirements. The nurses were happy with the system and bypassed the phones and overhead paging in preference of their headsets. However, during the system trial, the effect of noise on the automatic speech recognition (ASR) engine from sources like alarms, medical equipment, and wind noise from mobile users was an issue. Utterance tests were run using a variety of different Bluetooth headsets to measure the impact of various noise sources, S/N, and noise reduction algorithms on ASR performance. Common noise reduction methods were not helpful for the type of noise encountered, suggesting that solutions such as noise robust speech systems and noise reduction with adaptive microphone arrays currently being researched will be of great importance to future systems of this kind.

Constructions "apersonnelles" en latin

SummaryLatin grammars show a varied repertoir of the so called "impersonal" constructions. However under that label we find rather different phenomena exhibiting several regularity degrees. Together with real impersonal constructions, i.e., passive voice third person Singular forms, lacking subject, object and agent, other items such as weather and ethical verbs are added to the paradigm, though these conform a particular closed group that cannot be mirrored by others. Subjectless constructions formed by a third person plural active voice

Lenny Bruce without Tears

CAN YOU SHOUT Fuck! in a crowded nightclub? Maybe it would be all right if you were an etymologist, though professors ought not shout. You might avoid jail, oddly enough, if you said you were satirizing obscenity laws, or Fire Drills, or anything. But if you said it only in jest, for fun of it, watch out. Wear sneakers. Lenny Bruce's life in all its perversity, variety, tragedy, has so fascinated usread Albert Goldman's Boswellian biography-that it is difficult to think of him except in heroic cult terms: Icarus, Foul Mouth, Socrates, Drug Fiend, Defendant, Satirist. He is a totem figure carved of many different heads. It is time now to separate if we can Work from Life, and to try to know him for his real accomplishment, his real face-Comedian. An appropriate place to begin is Lenny Bruce's best joke and best sustained monologue of his career-the piece on Frank Dell's disastrous performance at Palladium Theatre. Bruce begins in his own voice, disclaiming concept of the room, geographic idea of happiness. But this skit deals with Frank Dell, Dean of Satire and Mimicry, who believes that his whole life would be meaningless if he did not play a class room. His agent tries to argue or humor him out of this obsession, but Frank is a driven man: Look, tired of playing toilets, man. I've had it. Of course it is a mistake. Frank is fine in lounges of Las Vegas for decent money-Is that spit? his agent argues. This agent, an insufficient creature in almost every way, does comprehend hard realities of business. His expertise is in dispassionate calculation of probabilities of supply and demand. His view of world is brutally limited, but he is absolutely right about Frank Dell. We are not encouraged to like agent-Lenny Bruce gives him busy Broadway agent's voice that is capable of sudden modulations from roughness into transparent motherly love: Sweetie baby bubby, sweetie. We do not like him, but he is right and Frank Dell is wrong. The music conductor at Palladium also knows immediately at rehearsal that Dell will fail. When he sees Frank's list of standard show biz impressions, he mumbles, Same crap week after week . . . disgusting. Frank apparently asks what he said, but conductor replies primly, I'm sure you do impressions different-you probably do them as children. The conductor, agent, and later house booker at Palladium, Val Parnell, all see Frank's limitations. They are bottom line-reality. They know exactly what will