Examining the Use of Intelligent Conversational Voice-Assistants for Improved Mobility Behavior of Older Adults in Smart Cities

Figures & data

Figure 1. Categorization of intelligent conversational voice-assistants.

Figure 2. Overview of the search process for the identified of sources.

Figure 3. Distribution of selected sources years distribution.

Figure 4. Distribution of employed methodology.

Figure 5. Distribution of country of selected sources.

Figure 6. Distribution of explored context in selected sources.

Figure 7. Distribution of disciple explored in selected sources.

Table 1. Related studies on the use of intelligent conversational voice-assistants for older people.

Author(s), year, and contributions	Explored research themes	Methodology employed	Research context	Countries
Arnold et al. (2022) explored the use of personal voice-assistants among older adults and associated outcomes.	Personal voice assistant Older adults Ageing in place	Literature review	Examined the degree of personal voice-assistants utilized for older adults, their technology readiness level.	Canada
de Arriba-Pérez et al. (2023) suggested an automated detection of cognitive impairment in older people.	Intelligent systems Monitoring Cognitive impairment	Experiment	Developed a system that can assess cognitive abilities of older people using machine learning algorithms.	Spain
Chen et al. (2021) investigated the barriers and design opportunities of using voice-assistants.	Gerontechnology Accessibility User experience design	Interviews	Provided insights regarding the issues for designing technologies for older adults, particularly related to healthcare.	United States of America
Duque et al. (2021) examined how to improve wellbeing of older people using digital voice-assistants.	Automation Conversational agents Design ethnography	Interviews	Indicated that digital voice-assistants aids older people in a dialectic relationship with automated content.	Australia
Gollasch and Weber (2021) researched on age-related variations preferences for using voice-assistants.	Voice interaction Conversational user interfaces	Survey questionnaire	Identified the particular age-related preferences of older people when they use a conversational user interface.	Germany
Mundhra et al. (2021) designed a humorous chatbot for senior citizens.	Chat-bot Information retention	Questionnaire	Assessed information retention in older people conversing with a chatbot.	Singapore
Schlomann et al. (2021) studies the capability and drawbacks of digital voice-assistants in older adults.	Voice-assistants Speech assistant Smart speaker	Literature review	Proposed detailed research designs to offer better insights into the adoption and use of voice-assistants in advanced age.	Germany
Valtolina and Hu (2021) developed a chatbot to improve the quality of life of older adults.	Conversational interfaces Chatbots usability Interaction design strategy	Interview	Described a chatbot discussing with older people, with age-related difficulties and the level of acceptability by older people.	Italy
Miura et al. (2020) deployed and assessed feedback feature for monitoring service for older people.	In-home long-term care Monitoring system Sensing agent	Experiment	Aimed to automatically provide feedback and guidance to improve the current mental state situation of older people.	Japan
Pradhan et al. (2020) examined the use of intelligent voice-assistants by older people with less technology use.	Conversational interfaces Voice-assistants Smart speakers	Daily diary entries, usage logs and interviews	Provided understanding on how older adults occasionally perceive and utilize voice-assistants .	United States of America
Ryu et al. (2020) designed a chatbot service to improve the mental health care of older people.	Interaction Mental health care User-oriented design	Field study Interview	Implemented a mental health care chatbot to helps in decreasing depression and anxiety among older people.	South Korea
Pradhan et al. (2019) explored the classification of smart speaker voice-assistants by older people.	Personification Ontology Voice-assistants	Interviews	Contributed how existing theories of anthropomorphism can be employed to voice-assistants.	United States of America
Sayago et al. (2019) investigated open issues faced in use of voice-assistants among older people.	Conversational user experience Voice-assistants design	Literature review	Studied adoption of voice-assistants for older people and explored perceptions and barriers to use of voice-assistants .	Spain, Australia, Ireland
Shalini et al. (2019) explored the design of voice-assistants to improve independence of older adults.	Voice-assistant Older adults Speech recognition	Focus group Test scenarios	Focused on human-computer design, mainly in voice-assistants to support older people to manage their health.	United States of America
Stigall et al. (2019) investigated older people use and perception of voice user interfaces.	Accessibility Voice user interface Natural language	Literature review	Presented the benefits of voice interaction for older people and barriers limited the adoption of this technology.	United States of America, Australia
Ho (2018) suggested a voice-controlled virtual assistants for older people faced with visual impairment.	Visual impairment Voice-controlled smart speakers	Literature review	Discussed on how virtual assistants assist the daily lives of older people with visual impairment to improves their wellbeing.	United Kingdom
Thakur and Han (2018) explored social acceptance of virtual assistants by older people.	Virtual assistants Social acceptance User experience	Sentiment analysis of tweets developed using RapidMiner	Utilized the concept of sentiment analysis to analyze the beliefs of older people use of virtual assistants.	United States of America
Shaked (2017) researched on virtual agents and avatars in relationship interfaces for the older people.	Decision making tools Virtual agents Home care monitoring	Literature review	Presented the challenge that impacts the design of interfaces can enable smooth interaction for the older users.	Israel
Tsiourti et al. (2016a) designed a framework as a multimodal virtual companion for older adults.	Artificial social companions Autonomous agent	Focus groups interviews Paper based surveys	Presented an architecture on the decision making and perception components which impacts affective behavior.	Switzerland Portugal Austria
Tsiourti et al. (2016b) designed a virtual assistive companion for older adults.	Conversational agents Agent architecture Multimodal interaction	Focus groups interviews Questionnaire	Carried out a user-centred design research that supported the design of the companion.	Switzerland Portugal Austria

Figure 8. Intelligent conversational voice-assistants for improved mobility behavior.

Table 2. Factors that impact use of intelligent conversational voice-assistants to supports older people.

Factors	Description
Ease of use	This factor is concerned with how easy and user friendly ICV is to older people when they navigate the auditory interface. This factor assessed if it requires a lot of learning effort for older people to use ICV to support mobility services. This requirement is necessary to accommodate the older people physical and cognitive abilities.
Visualization	This factor is concerned with the external appearance or avatar of the ICV within the device being deployed whether the avatar is a human-like agent, an image an animal, or just an application user interface. This requirement assesses how likeable the ICV is and how well it characterizes or personalized the mobility needs of older people user (Shaked, 2017).
Personalization	The personality of the Persona or the Avatar is conveyed via speech, visual, facial, and possible emotional gestures can influence the use of conversational voice-assistants. Personalization relates to how ICV will be able to provide different services particularly personalized to the daily mobility needs and walkability tasks (Shaked, 2017).
Availability of analytics	This factor plays a crucial role in achieving personalization. The provision of analytics facilitates the establishing of user mobility patterns and this offers predictions for future mobility behavior for older users.
Error recovery	This factor involves the capability of ICV to easily recover from a dialogue misunderstanding or mistake rather than creating an infinite conversation loop and an unpleasant interaction experience (Shaked, 2017). It is important to create a smooth interaction with older people especially in a dialogue for wayfinding and walkability (Shaked, 2017). Error recovery is challenging to avoid and requires use of data analytics technologies integrated with NLP module.
Latency	This factor is important to improve user experience as the delay response between input into the system and the desired reply or output (Barros Pena et al., 2021a). Latency highly affects the quality of the interaction as well as the dialog flow. Thus, high latency which is the time taken in seconds per interaction may be observed in by older people as lack of ability to understand him/her and this create a sense of misunderstanding (Shaked, 2017).
Integration	ICV should possess the ability to incorporate different input/output technologies into the dialogue that enables older people to choose which auditory interface is preferable to accommodate the older users’ capabilities.
Gamification	This factor involves the inclusion of gaming functionality and rewards for task execution is highly recommended for as it supports the continuity and familiarity of ICV with older people (Shaked, 2017).
Appearance	There is need for ICV with a realist and a friendly voice and face that makes older people feel comfortable. Findings from the literature suggested that availability of different virtual characters, with different looks base on different age and gender, realistic human-like expression is preferred over cartoonlike appearance and formal as compared to informal look (Tsiourti et al., 2016a).
Verbal and non-verbal capabilities	This factor involves the provision of a ICV that has seamless communicative proficiencies, such as the capabilities to perceive and interpret vocal and non-audio input (for example verbal language and facial expressions), and to express voiced and non-vocal behavior (Tsiourti et al., 2016a).
Decision-making and personality	The ICV should show emotional responses and coherent behavioral that can be understand by older people as indication of natural and social personality (Tsiourti et al., 2016a).
Perception and awareness	ICV should be able to perceive their deployed environment once its operational and it should be aware of the current location/condition of the older people. Older people envisioned a ICV that can understand their emotional state (e.g., sadness, joy, etc.) and activity (e.g., walking, idle) and utilize this data to adapt its behavior to offer mobility and walkability support.
Safety	ICV should be able to monitors older people and recognizes abrupt human body motions that indicate instability or a sudden fall or a call for help, based on a predefined voice command. In the case of an emergency, it should initiate a dialog to acknowledge the detection and reassure the user and if necessary, automatically dispatches a phone call to a designated caregiver or emergency center (Tsiourti et al., 2016a).
Guidance	Since older people are disposed to forgetfulness, it is expected that ICV can recognize emergency situations (e.g., injury, falls), and help older people detect objects such as signs, symbols, etc. in the build environment utilizing real-time vision-based detection of prior learned objects, when possible (i.e., when a physical object is in the area of view of the phone camera) (Tsiourti et al., 2016a).
Privacy	Furthermore, in accordance with General Data Protection Regulation (GDPR) other associated risk factors that impact use of ICV relates to the perceived threat to user’s privacy due to the data that is collected by the system based on the individual’s information (Chen et al., 2021; Hoy, 2018; McLean & Osei-Frimpong, 2019; Pradhan et al., 2020). Although ICV provide benefits to older people, the adoption of digital technologies such as AI based machine learning can pose threats to users’ privacy. As there is a possibility for data to be stolen or leaked (Hoy, 2018).
Security risk	Findings from the literature stated that ICV such as the Amazon Echo have security risk that can be manipulated by individuals with unauthorized access. ICV can execute high priority tasks utilizing private account details, assist in making appointments, look up available travel service information, reserve mobility modes on behalf of older people, and make payments on behalf of user and the payment details can be faced with security issues.
Data access	Another issue with ICV is data security. Anyone with access to ICV based system can query it, get information about the personal accounts and digital services connected with the device, and query it to perform tasks. This can be a security risk as these devices will read out emails, SMS, calendar contents, and other personal information (Stigall et al., 2019). To address this type of unauthorize data access Google has lately upgraded its voice assistant software to incorporate voice printing, which distinctively identifies each user using their voice to stop the device from reading out personal information to unauthorized voices. Apple is also training Siri to recognize an individual’s voice, whereas Amazon’s Alexa is working to deploy a similar voice printing module. There are now options to set a voice activated passcode to confirm user accounts (Hoy, 2018). Another approach to address access concern in ICV all collected data should be destroyed after analysis or no data should be recorded (Stigall et al., 2019).

Figure 9. Requirement specifications for designing conversational voice-assistants.

Table 3. Requirement specifications and strategies for designing intelligent conversational voice-assistants for older people.

Methods	Requirements	Description and user-centered strategies
Ability-based design framework (Chen et al., 2021; Wobbrock et al., 2011; 2018)	Ability	Re-enforces the need to focus on the ability and not on the dis-ability, thus striving to capitalize on the capabilities of older users to augment what they need to do. Older people should ubiquitously access mobility availability and accessibility information using auditory interface without requiring many efforts and time when they are walking in cites. The system should also allow older people to easily change pre-configured settings based on their mobility and walkability need (e.g., enable/disable ICV connected to a specific feature in their device). The provision of avatars or visual output for enhanced social communication and user interaction experience. Availability of multimodal output to support mobility when older people with sensory impairment use these systems.
	Accountability	Encourages system designers to improve performance by changing the systems and not trying to change older users, thus it is recommended to leave older users as they are. ICV needs to accurately understand older people intent and provide procedure to restore interaction upon system failures when in use by older people walking in cities or in transit in public transportation.
	Adaptation	The auditory interfaces may be user-adaptable or self-adaptive to offer the best possible match to suitable to older people mobility needs and ability. Personalized response should be based on older people mobility interests and hobbies. The provision of personalized on-demand active mobility and safe walking advice.
	Transparency	The auditory interface can offer users awareness of any adaptations and the resources to check, reverse, reject, change back, preview, store, retrieve, and pre-test those adaptations for improve mobility recommendations. Older people should have easy, clear, and plain guidance during system troubleshooting and setting up phases.
	Performance	The systems can monitor users’ usage performance by logging and predicting measures to improve performance usage. ICV should be able to sense and compute older people mobility pattern to provide personalized guidance.
	Context	The systems should be able to sense context or environment when in use and propose recommendation to improve mobility and walking for older people based on their existing ability. Older people should be able to plan their mobility needs and adapt to the pedestrian environment using ICV.
	Commodity	The deployed system should consist of readily available, inexpensive, low-cost device hardware and software. The integration of ICV into current technologies should not require high cost of ownership for older people.
	Privacy and trust	The system should employ privacy by design measures during the development for improve security and privacy of older people data and to increase user trust of the technology. Employ measures to guarantee user confidentiality mostly for contextual data and unintended speech, as well as to improve the trust between older people and ICV.
ISO 9241-110 ergonomics of human-system interaction (Gollasch & Weber, 2021; ISO, 2020)	Suitability for the task	The speech recognition module should be able to recognize different group of older people (between the age of 65–74, 75 and older), to address usability issues faced by older people amidst use of these ICV. The speech recognition module should be able to understand indistinct accents or pronunciation (this may be due to the local dialects, idiom or speech impairment faced with older people between the age of 65 and older. It should possess the capability to handle and process repetitions, pauses in speech, and incomplete words. Use simplified activation word and possess the capability to process different languages if possible translation
	Self-descriptiveness	The system should be able to manage the presentation of several information to older people by being able to cope with complex conversations using simple step-by-step dialogues during movement. Since the system is intended for older people it is recommend incorporating simplicity in the sentence structure with short sentences with few statements per sentence. The system should be able to understand single question in each dialogue, preferably, yes/no questions and to some extend be able to decode complex questions. The system should signal user request to speak (as input phase) with limited marginal error.
	Conformity with expectations	Involves high expectations from users who perceived the interaction with ICV to be human like dialogue. The system should be able to cope with change of topic, perceive topic switches, and ambiguous communication in different conversation context such as in mobility and walkability/wayfinding. The possibility to show variation in the construction of output statements, rephrasing of similar or same statements. Reiterating statements in combination with common, renewed, and rephrasing reference to the current discourse of mobility and walkability with older people. There should be balance with the speech speed and clarity. Thus, speaking slowly and clearly, using frequent tone of voice, thereby optimizing voice clarity over speed Better sensitivity for recognizing the activation word and always reacting to the activation word, even for older people with unclear accents or pronunciation.
	Learnability	Involves inclusion of precautionary questions as safety check to mitigate unplanned irreversibly actions. Make recommendations that proactively improves the personalization. Provide help with directions and support that proactively improve mobility and safe walkability/wayfinding.
	Controllability	Provision of options to change the voice or audio (volume, pitch, type, speaking rate, etc.), as well as personality and customizability of pronunciation. Prevent unintentional interruption of the conversation when in use not to compromise safe mobility.
	Error robustness	Employing checking procedure to assess what has been understood by the user by prompting for repeat entered information. In case of an error message the level of information provided as an error message should be less detailed. In case of any unforeseen misunderstandings the system should prompt user and ask for repetition when needed.
	User engagement	The system should preferably use formal communication in relation to informal communication and provide positive feedback to improve the mode of older people. Also, personal communication is preferred as compared to neutral communication. When the user introduces themselves to ICV, the system should provide an introduction of itself back to the user to build trust.

Arnold, A., Kolody, S., Comeau, A., & Miguel Cruz, A. (2022). What does the literature say about the use of personal voice-assistants in older adults? A scoping review. Disability and Rehabilitation. Assistive Technology, 19(1), 100–111. https://doi.org/10.1080/17483107.2022.2065369

 PubMed Web of Science ®Google Scholar

de Arriba-Pérez, F., García-Méndez, S., González-Castaño, F. J., & Costa-Montenegro, E. (2023). Automatic detection of cognitive impairment in elderly people using an entertainment chatbot with natural language processing capabilities. Journal of Ambient Intelligence and Humanized Computing, 14(12), 16283–16298. https://doi.org/10.1007/s12652-022-03849-2

 Google Scholar

Chen, C., Johnson, J. G., Charles, K., Lee, A., Lifset, E. T., Hogarth, M., Moore, A. A., Farcas, E., & Weibel, N. (2021). Understanding barriers and design opportunities to improve healthcare and QOL for older adults through voice-assistants [Paper presentation]. In The 23rd International ACM SIGACCESS Conference on Computers and Accessibility (pp. 1–16).

 Google Scholar

Duque, M., Pink, S., Strengers, Y., Martin, R., & Nicholls, L. (2021). Automation, wellbeing and digital voice-assistants: Older people and google devices. Convergence: The International Journal of Research into New Media Technologies, 27(5), 1189–1206. https://doi.org/10.1177/13548565211038537

 Web of Science ®Google Scholar

Gollasch, D., & Weber, G. (2021). Age-related differences in preferences for using voice-assistants. In MuC ‘21: Proceedings of Mensch und Computer 2021 (pp. 156–167). https://doi.org/10.1145/3473856.3473889

 Google Scholar

Mundhra, R., Lim, T. J., Duong, H. N., Yeo, K. H., & Niculescu, A. I. (2021). Towards a humorous chat-bot companion for senior citizens. In Conversational dialogue systems for the next decade (pp. 31–39). Springer.

 Google Scholar

Schlomann, A., Wahl, H.-W., Zentel, P., Heyl, V., Knapp, L., Opfermann, C., Krämer, T., & Rietz, C. (2021). Potential and pitfalls of digital voice-assistants in older adults with and without intellectual disabilities: Relevance of participatory design elements and ecologically valid field studies. Frontiers in Psychology, 12, 1–5. https://doi.org/10.3389/fpsyg.2021.684012

 Web of Science ®Google Scholar

Valtolina, S., Hu, L. (2021). Charlie: A chatbot to improve the elderly quality of life and to make them more active to fight their sense of loneliness. In CHItaly 2021: 14th Biannual Conference of the Italian SIGCHI Chapter (pp. 1–5.).

 Google Scholar

Miura, C., Saiki, S., Nakamura, M., & Yasuda, K. (2020). Implementing and evaluating feedback feature of mind monitoring service for elderly people at home [Paper presentation]. Proceedings of the 22nd International Conference on Information Integration and Web-Based Applications & Services (pp. 390–395). https://doi.org/10.1145/3428757.3429124

 Google Scholar

Pradhan, A., Lazar, A., & Findlater, L. (2020). Use of intelligent voice-assistants by older adults with low technology use. ACM Transactions on Computer-Human Interaction, 27(4), 1–27. https://doi.org/10.1145/3373759

 Web of Science ®Google Scholar

Ryu, H., Kim, S., Kim, D., Han, S., Lee, K., & Kang, Y. (2020). Simple and steady interactions win the healthy mentality: Designing a chatbot service for the elderly. Proceedings of the ACM on Human-Computer Interaction, 4(CSCW2), 1–25. https://doi.org/10.1145/3415223

 Google Scholar

Pradhan, A., Findlater, L., & Lazar, A. (2019). Phantom Friend" or" just a box with information" personification and ontological categorization of smart speaker-based voice-assistants by older adults. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), 1–21. https://doi.org/10.1145/3359316

 PubMedGoogle Scholar

Sayago, S., Neves, B. B., Cowan, B. R. (2019). Voice-assistants and older people: Some open issues. In Proceedings of the 1st International Conference on Conversational User Interfaces (pp. 1–3.).

 Google Scholar

Shalini, S., Levins, T., Robinson, E. L., Lane, K., Park, G., & Skubic, M. (2019). Development and comparison of customized voice-assistant systems for independent living older adults. In International conference on human-computer interaction (pp. 464–479). Springer.

 Google Scholar

Stigall, B., Waycott, J., Baker, S., Caine, K. (2019). Older adults’ perception and use of voice user interfaces: A preliminary review of the computing literature. In Proceedings of the 31st Australian Conference on Human-Computer-Interaction (pp. 423–427.).

 Google Scholar

Ho, D. K. (2018). Voice-controlled virtual assistants for the older people with visual impairment. Eye (London, England), 32(1), 53–54. https://doi.org/10.1038/eye.2017.165

 PubMed Web of Science ®Google Scholar

Thakur, N., & Han, C. Y. (2018). An approach to analyze the social acceptance of virtual assistants by elderly people [Paper presentation]. Proceedings of the 8th International Conference on the Internet of Things (pp. 1–6). https://doi.org/10.1145/3277593.3277616

 Google Scholar

Shaked, N. A. (2017). Avatars and virtual agents–relationship interfaces for the elderly. Healthcare Technology Letters, 4(3), 83–87. https://doi.org/10.1049/htl.2017.0009

 PubMedGoogle Scholar

Tsiourti, C., Quintas, J., Ben-Moussa, M., Hanke, S., Nijdam, N. A., & Konstantas, D. (2016a). The CaMeLi framework—A multimodal virtual companion for older adults. In Proceedings of SAI intelligent systems conference (pp. 196–217). Springer.

 Google Scholar

Tsiourti, C., Ben Moussa, M., Quintas, J., Loke, B., Jochem, I., Lopes, J. A., & Konstantas, D. (2016b). A virtual assistive companion for older adults: Design implications for a real-world application. In Proceedings of SAI intelligent systems conference (pp. 1014–1033). Springer.

 Google Scholar

Barros Pena, B., Clarke, R. E., Holmquist, L. E., Vines, J. (2021a). Circumspect users: Older adults as critical adopters and resistors of technology. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (pp. 1–14). https://doi.org/10.1145/3411764.3445128

 Google Scholar

Hoy, M. B. (2018). Alexa, Siri, Cortana, and more: An introduction to voice-assistants. Medical Reference Services Quarterly, 37(1), 81–88. https://doi.org/10.1080/02763869.2018.1404391

 PubMedGoogle Scholar

McLean, G., & Osei-Frimpong, K. (2019). Hey Alexa… examine the variables influencing the use of artificial intelligent in-home voice-assistants. Computers in Human Behavior, 99, 28–37. https://doi.org/10.1016/j.chb.2019.05.009

 Web of Science ®Google Scholar

Wobbrock, J. O., Kane, S. K., Gajos, K. Z., Harada, S., & Froehlich, J. (2011). Ability-based design: Concept, principles and examples. ACM Transactions on Accessible Computing, 3(3), 1–27. https://doi.org/10.1145/1952383.1952384

 Google Scholar

Wobbrock, J. O., Gajos, K. Z., Kane, S. K., & Vanderheiden, G. C. (2018). Ability-based design. Communications of the ACM, 61(6), 62–71. https://doi.org/10.1145/3148051

 Web of Science ®Google Scholar

ISO. (2020). ISO 9241-110:2020 –Ergonomie der Mensch-System-Interaktion [Ergonomics of human-system interaction-Part 110: Interaction principles]-Teil 110. Interaktionsprinzipien.

 Google Scholar