Technological and medical advances have led to a demographic shift in the population, with the number of older adults constantly increasing. According to the United Nations [1], older adults (aged 60-65 years) will represent 16% of the world’s population in 2050. In addition, life expectancy is increasing, from 64.2 years in 1990 to 72.6 years in 2019, and is expected to reach 77.1 years in 2050 [1]. However, there is a wide diversity of health conditions among older adults. The health status of older adults is dependent on multiple factors, including nonmodifiable genetic factors and environmental factors, such as lifestyle [2]. Thus, older adults represent a very heterogeneous population with multiple and diverse needs and desires. With advancing age, the loss of functional independence; frailty; and other health diseases such as cardiovascular problems, cancers, osteoarthritis, osteoporosis, or major neurocognitive disorders may appear [3-5]. Among age-related conditions, major neurocognitive disorders (eg, Alzheimer disease) receive particular attention due to the increasing prevalence of these diseases [6].

The aging population is not only a public health issue but also a socioeconomic one. To face this challenge, it is important to develop preventive measures to support active and healthy aging and to preserve the independent functioning and quality of life of older adults. The adoption of healthy behaviors can help prevent or delay the onset of pathologies or treat them if detected early [7].

Preventive health measures can be supported through new technologies, such as robotic coaching solutions (RCSs) that promote healthy aging among older adults [8,9]. RCSs have been defined as personalized systems that continuously monitor the activities and environment of the user and provide them with timely health-related advice and interventions [10-12]. These systems can help users define and achieve different health-oriented goals [12].

RCSs may encompass artificial intelligence (AI) technologies that can analyze user data, personalize coaching programs, and adapt recommendations based on each individual’s needs [1,13-19]. RCSs can involve robots equipped with sensors such as cameras, microphones, or motion sensors to collect real-time data about the user, AI, and programming that enables their interaction with users [20,21]. These technologies are often equipped with voice and visual recognition and learning capabilities [20,21]. They can benefit from advanced natural language processing techniques, which allow for understanding of the user’s input, facilitating natural and effective communication [22]. RCSs can offer guidance, support, and feedback based on preprogrammed information or real-time data analysis. These data can inform coaching strategies and allow RCSs to provide users with relevant feedback [8].

RCSs can also encompass a virtual agent, which refers to a computer program or an AI system that interacts with users in a manner that simulates human conversation [14,18,23]. A virtual agent is an animated character capable of adopting a social behavior mimicking that of humans to encourage the users to make changes in their habits [14]. Virtual agents might take the form of a chatbot, voice assistant, or other AI-driven communication system [14]. Biometric monitoring devices to track physiological data such as heart rate, sleep patterns, or stress levels can also be included in RCSs [8,20,21]. These data can contribute to the configuration of personalized coaching plans. RCSs can also encompass advanced data analytics that can process large data sets generated by users’ interactions and behaviors. This functionality helps in identifying patterns, trends, and areas for improvement in coaching strategies [24]. Integrating Internet-of-Things devices in RCSs can provide additional data points about a user’s environment, lifestyle, or habits, thus contributing to a personalized coaching approach [25].

Health-oriented RCSs could enable users to lead a healthy lifestyle, by identifying needs and goals and providing appropriate risk predictions and individualized recommendations [12,26-28]. There are RCSs dedicated to a particular domain, such as physical activity or motor rehabilitation [9,16]. Others may have the objective of promoting independent and healthy aging [29].

Promoting active and healthy aging can allow older adults to maintain their independence and continue to live at home [4,30], which is a wish of many [3]. This intervention could also help to reduce the need for assistance, usually provided by informal caregivers and health professionals [4,19,30-33]. Furthermore, RCSs could lead to a reduction in individual and collective health care expenses [4,32,34] by easing access to health and social care interventions to a wide population, including hard-to-reach (eg, geographically isolated) individuals. However, although the use of health-related RCSs could have many benefits, several ethical issues arise with their development and implementation in human environments [3,35-38].

For RCSs to contribute to active and healthy aging, it is important that all the stakeholders (engineers, geriatricians, psychologists, etc) involved in their design and implementation refer to an ethical framework [3,38]. It is also important to inform society (politicians and legal experts) about such an extension of technology in people’s lives (private, professional, medicosocial, and commercial context), so that we can create a legal framework for the use of these technologies. An analysis of the way in which ethical and legal dimensions have been addressed by studies, in the field of RCSs for health care, seems useful to support the key actors in their development and implementation. The growing interest in the ethical questions associated with the use of social and assistive robots is evidenced by the volume of literature reviews [3,12,18,31,32,37,39-51] on the topic.

Now, it appears appropriate to systematically examine this body of work, focusing on the ethical analysis, and provide an overview of the literature. Therefore, we performed a review of the literature on RCSs for older adults using the European Network of Health Technology Assessment (EUnetHTA Core Model; version 3.0) model [52] for analysis. This Health Technology Assessment (HTA) model makes it possible to assess the intended and unintended consequences of the use of a specific technology regarding multiple domains (eg, technological, ethical, clinical, and organizational), providing methods and concepts for this analysis [53]. Therefore, HTA is a process that informs decision-making about the introduction of new technologies such as RCSs in health care. It also seems necessary to issue guidelines for the development and implementation of health-oriented RCSs [54].

The objective of this study was to highlight the main ethical questions and corresponding recommendations linked to the use of RCSs for older adults for engineers, researchers, and health professionals in this field. For this purpose, we conducted a narrative literature review using the ethical dimension of the EUnetHTA Core Model to guide the analysis. To the best of our knowledge, such a study has not been conducted so far.

A thematic analysis of the literature was performed to identify publications that describe RCSs for supporting older adults in health care and prevention and those that address ethical issues and recommendations regarding their development and implementation. The methodology used for the narrative review was inspired by the study by Green et al [55].

The review encompassed papers focusing on all populations, with particular attention to older adults. It focused on the concept of RCSs for health, while also incorporating publications discussing other health technologies for older adults if the authors have delved into relevant ethical considerations for their development or implementation.

The context of the review revolved around the use of RCSs (or related technologies), especially for older adults, across diverse living environments such as homes, hospitals, and nursing homes. Publications addressing RCSs and related ethical issues within the health care domain were considered, whereas those focusing solely on technical aspects (eg, AI and deep learning) or those outside the health care domain were excluded.

Various types of publications, including theoretical papers, research studies, and reviews, were included if they offered ethical reflections or recommendations for RCS use in health care. These reflections and recommendations were expected to align with the topics and issues of the ethical dimension of the EUnetHTA Core Model.

All publications, regardless of language (English or French), were eligible if published after 2011. This time frame was chosen considering the technological advancements over the past decade, which may have influenced the evolution of ethical issues and recommendations in the field of remote care systems and related technologies. Textbox 1 summarizes the inclusion and exclusion criteria adopted for the selection of papers in this review.

The review was conducted using the following keywords: “seniors,” “older adults,” “social robots,” “assistive robots,” “assistive technology,” “robots,” “virtual coach,” “e-coaching,” “coaching system,” “coaching device,” “ethics,” and “recommendations.”

The review was performed between August 2021 and December 2021 using the PubMed, CINAHL, Embase, Scopus, Web of Science, IEEE Explore, SpringerLink, and PsycINFO databases.

This search allowed us to find 4928 initial publications. Then, secondary research using references from other articles and the same inclusion criteria was conducted. This search allowed us to find 13 additional papers.

In total, 4943 papers were analyzed. The selection of the final publications was performed after reading the title and abstract first and, then, the full article. This selection process helped us to exclude irrelevant papers and duplicates (Figure 1). In total, 0.51% (25/4943) of the papers were included in our review.

The selected papers were analyzed using the ethical domain of the EUnetHTA Core Model [52]. Proper registration of the use of EUnetHTA Core Model for the purpose of this review was made on the HTA Core Model website [52].

The model was developed for the production and sharing of HTA information, allowing for the support of evidence-based decision-making in health care, but it can also be customized to other research needs. The EUnetHTA Core Model is composed of 9 domains, each including several topics. Each topic also includes different issues (ie, questions that should be considered for the evaluation of health technologies). Thus, the model is structured into 3 levels: domain (level 1), topic (level 2), and issue (level 3). The combination of a domain, topic, and issue is linked to an assessment element ID, which can be identified using a specific code for standardization purposes (B0001, B0002, etc).

The main EUnetHTA model domains include the following: (1) health and current use of the technology, (2) description and technical characteristics of the technology, (3) safety, (4) clinical effectiveness, (5) costs and economic evaluation, (6) ethical aspects, (7) organizational aspects, (8) patient and social aspects, and (9) legal aspects.

The ethical domain (level 1) in the EUnetHTA Core Model [52] includes 5 topics (level 2): “benefit-harm balance,” “autonomy,” “respect for people,” “justice and equity,” and “legislation.” Each of these topics includes several issues (level 3) [52].

In this study, 2 authors (CP and ASR) independently analyzed the 25 selected articles. First, they read the articles several times to improve familiarity with the ideas addressing the ethical aspects of RCSs. Then, in each publication (methods, results, and discussion sections), they identified segments of data that were relevant or captured an idea linked to the “ethical” domain of the model. A subsequent exploration of the coded data (sentences or set of statements) was performed to get a more precise classification at the topic level (level 2) and at the issue level (level 3). Then, the coding was performed using the HTA nomenclature. The 2 experts (CP and ASR) compared their results. In a few cases, the coding results showed a lack of consensus between the 2 coding authors, which was resolved through a subsequent discussion between them. Interrater correlation was not calculated.

A thematic analysis using the EUnetHTA framework for conducting a literature review has been described in other studies [56,57]. Furthermore, the use of EUnetHTA to perform an ethical analysis of health technologies has already been proposed [58]. The 25 selected articles were all coded using this methodology. Some authors have previously emphasized the possibility of overlapping issues between topics in the HTA analysis. They have suggested to assess the overlapping issues in the most relevant topic section [59].

This review was not registered, and a protocol for the review was not prepared.

This section aims to summarize the ethical analysis performed regarding the use of RCSs with older adults and to provide recommendations for ethical use of these devices. Table 1 presents a synthetic summary of the elements presented in this section.

The use of technologies, such as RCSs, in the health care field has grown significantly in recent years [17,18]. RCSs are increasingly being used for older adults with the aim of promoting healthy behaviors, quality of life, and well-being. However, the use of RCSs also raises several ethical challenges regarding the cost-to-benefit balance of these new care practices, respect for the autonomy of users, respect for privacy, justice and equity linked to their access, or need for a suitable legal framework. Such challenges could be addressed by establishing relevant recommendations for the development and use of RCSs. Some guidelines regarding the use of robotic systems have been published [49,83]. Moreover, in April 2021, the European Commission unveiled the first legal framework about AI [84]. However, to the best of our knowledge, no recommendations have been proposed in this field directly linked to an analysis of the literature dealing specifically with these ethical issues and potential solutions to address them.

This narrative review identified 25 articles in which authors highlighted ethical issues and recommendations related to the use of RCSs and similar technologies. The use of the EUnetHTA Core Model for the analysis of these articles made it possible to classify the information retrieved in the publications according to 5 main ethical topics—“benefit-harm balance,” “autonomy,” “respect for persons,” “justice and equity,” and “legislation”—and to provide a detailed analysis of RCS-related ethical issues. Our review also aimed to identify recommendations for better development, diffusion, and use of RCSs by a population of older adults.

Technology devices, such as RCSs, are used with older adults to enable them to live independently; to enhance their quality of life and well-being; and, therefore, to cope with the increasing care demands for older populations. RCSs may be used to encourage a range of health-related goals: physical, cognitive, nutritional, social, and emotional domains. To be effective, RCSs must be able to motivate the user by providing highly personalized care programs [85,86]. However, studies have shown that not all potential target users are included in the development of these devices [37,87,88]. Therefore, RCSs design might fail to meet a wide range of users’ needs, capabilities, and wishes. Thus, it is essential to apply “user-centered design” approaches and involve target users with various sociodemographic characteristics and technology experience throughout the development process. A strong involvement of the intended users of these systems in their design process would also improve the quality of the information provided to potential users of RCSs regarding their operation, type of data collected, and potential benefits of the technology. In this way, the involvement of the users would improve the quality of the process of obtaining the consent required from older adults to use the technology.

Another ethical challenge related to the use of RCSs is the fact that their wide implementation for older adults’ care may affect the distribution of health care resources. For instance, it has been found that for some older adults and informal and formal caregivers, the use of RCSs could replace humans in many caregiving tasks, eventually leading to a suppression of jobs or to a degradation of the quality of health care services [17,34,48,61]. In this regard, the participation of a third person (professional, volunteer, or family member) as a “human coach” could be considered when implementing RCSs in the older adults’ environment. This “human coach” could help build a “chain of trust” by being an intermediary between the RCS and the user. On the one hand, the involvement of a real person in the use of the RCS could reduce the risk of replacement of human assistance by technological assistance. On the other hand, the “human coach” could help enhance the acceptability and usability of the device, while at the same time, reassuring the user and providing recommendations to the developers, so that the RCS is consistent with users’ needs and desires. However, the benefits of involving a “human coach” in the RCS service provision has yet to be evaluated by scientific studies.

According to some studies [3,39,41,51,65], the use of RCSs can have an impact on social relationships, reducing human contact and even altering social relationships by creating tension between older adults and their caregivers. Thus, it would be interesting to identify the repercussions and implications of these devices in older adults’ daily life and in the life of the members of their social environment through new studies. It also seems necessary to evaluate the organizational impact of the implementation of RCSs and to identify potential obstacles to their use in the care professionals’ work context.

Our analysis also confirmed that for RCSs to provide personalized health-related recommendations, the collection of sensitive data is necessary. Data collection in this context also raises several ethical issues. For instance, personal data can be exposed to hacking and misuse. Proper data management, anonymization, and encryption are essential to protect the personal data of RCS users [86]. In addition, researchers and developers in this field must evaluate RCSs before implementation to ensure that they do not cause physical or moral harm to users. Thus, it has been suggested that stakeholders refer to local and regional regulatory and safety standards to guide their development and use.

Finally, our analysis also discussed how legal and ethical frameworks regarding the use of RCSs need to be adapted to cope with the constant development of new technologies. So far, existing legal frameworks are not yet adequate to respond effectively to the question of liability in case of damage caused by RCSs, particularly because these devices are becoming increasingly autonomous [80]. The establishment of “operational ethics committees in digital sciences and technologies” could help in the development and conduct of projects in this area [60]. Guidelines should be established to identify the types of applications and technological devices that require regulatory review and approval [44]. Research projects and working groups involving users, researchers, and lawyers should be set up to further investigate the legal and ethical issues related to the use of RCSs.

Some countries and regions, such as Europe and Japan have initiated the work of structuring relevant legal and ethical frameworks; however, their orientations and measures may differ culturally [78]. Future studies in the area of RCSs could consider the influence of cultural and socioeconomic specificities of the contexts of experimentation (countries and regions) regarding the acceptance and use of RCSs by older adults and formal and informal caregivers and regarding the definition of ethical and legal frameworks governing their uses. Therefore, the use of validated and widely applied analysis frameworks, for example, the Western, Educated, Industrialized, Rich and Democratic framework [89], formulated to measure countries’ commonalities in their approaches to the interpretation of behavioral research findings (eg, regarding technology adoption) could be interesting. The Western, Educated, Industrialized, Rich and Democratic framework [89] could help not only to explore the differences among countries regarding the validation and adoption of new technologies for older adult care but also to seek greater cultural and demographic diversity in technology research.

This dimension of cross-cultural comparison has received particular attention in the framework of a current international research partnership between Europe and Japan, such as the EU-Japan Virtual Coach for Smart Ageing (e-VITA) project. This project aims to develop a cross-cultural RCS that can be tailored to the needs of healthy older adults to promote aging well. The e-VITA RCS will be made available to older adults in their homes, which raises many of the ethical questions discussed in this paper. Therefore, the study will require the researchers to set up procedures adapted not only to the users but also to the 2 cultures (European and Japanese), respecting the corresponding ethical and legal regulations. Thus, it would be interesting to perform an analysis of the ethical issues raised by users from different countries and cultures within the framework of the e-VITA project.

A narrative review of the literature was conducted to provide a nonexhaustive synthesis of the various ethical concerns and recommendations when using RCSs for older adults. This review has some limitations. Only articles in French and English were included. Some articles indicating ethical concerns or recommendations may not have been included when this information was not mentioned in the keywords or abstract.

The use of RCSs in the context of health care, particularly with an older adult population, tends to show many benefits. RCSs have the potential to improve the quality of life of older adults and their independence. When used in an ethical and appropriate manner, RCSs can help improve older adults’ emotions and cognitive and physical abilities and promote social relationships. By helping older adults to continue living at home for as long as possible, the use of health-oriented RCSs could help to address some of the challenges resulting from demographic aging. However, the use of these new health care technologies involves some ethical concerns, with the most cited issues being not only the risk of accidents, lack of reliability, loss of control, risk of deception, and risk of social isolation but also the confidentiality of data and liability in case of safety problems.

Some recommendations have been made in the past regarding the use of social and assistive robotic technologies for older adults, such as considering the opinion of target users; collecting their consent; training the care professionals to use them; and ensuring proper data management, anonymization, and encryption. However, the integration of RCSs in current health practices and, particularly, in the private homes of older adults can be disruptive. It requires the establishment of scalable and adapted ethical and regulatory frameworks that follow the technology progress and the social and digital change of society Thus, studies are needed to identify new ethical concerns arising from the organizational impact of the implementation of RCSs in different contexts, especially in the homes of older adults. The influence of cultural and socioeconomic specificities of the contexts of experimentation (countries and regions) regarding the acceptance and use of RCSs by older adults and formal and informal caregivers is also an area of interest for future studies.