Ethical considerations in home monitoring technologies for persons living with cognitive impairment: a scoping review

Abstract

Background and Objectives

As the global population of people living with cognitive impairment grows, Home Monitoring Technologies (HMTs), such as cameras, motion sensors, wearable trackers, and artificial intelligence enabled ones are increasingly used to enhance safety and support aging in place. However, these technologies raise ethical concerns, particularly regarding privacy, autonomy, trust, and transparency. This scoping review explores these ethical implications and identifies key themes to inform future research, practice, and policy development.

Research Design and Methods

Following Arksey and O’Malley’s scoping review framework, systematic searches were conducted in PubMed, EMBASE, CINAHL, and PsycINFO (Arksey & O’Malley (2005). Scoping studies: Towards a methodological framework. International Journal of Social Research Methodology, 8, 19–32). Studies were included if they examined HMTs for people living with cognitive impairment and addressed ethical concerns. Our eight central themes were derived inductively during data synthesis, and the Rubeis’ 4D Risks Framework offered a valuable conceptual scaffold to organize and interpret the broader patterns of ethical risk.

Results

A total of 110 publications from 30 countries were reviewed. Ethical concerns were identified in each of the 4 areas of the framework, including privacy violations, loss of autonomy, erosion of trust, and unintended consequences such as social isolation and reduced human interaction. Person-centered design approaches, which engage both people with cognitive impairment and caregivers, were identified as crucial for mitigating risks and fostering ethical implementation.

Discussion and Implications

Findings underscore the need for evidence-informed guidelines that explicitly incorporate ethical frameworks to ensure consideration of the balance of health and safety with autonomy and dignity.

Cognitive impairment, including Alzheimer’s Disease (AD), is a growing global concern. In the U.S., 6.7 million older adults are currently living with AD, a number projected to reach 13.8 million by 2060 (Alzheimer’s Association, 2023). Globally, the population of individuals living with AD is projected to grow to 131.5 million by 2050 (Dong et al., 2019). Ratnayake et al. (2022) found that 71% of older adults aged 50 years and older expressed that aging in place allowed them to maintain independence, dignity, and quality of life in their own homes. While aging in place is a practical and deeply personal choice for many older adults, balancing their independence with safety is a critical challenge. Persons living with cognitive impairment are particularly vulnerable to risks such as illness, confusion, and accidents, which further complicate caregiving efforts. However, a projected 33% shortage in home health aides threatens the availability of professional support, placing increasing pressure on family caregivers. Currently, 11 million informal caregivers provide an estimated 16 billion hours of unpaid care annually, often at great personal cost (Alzheimer’s Association, 2023; Horovitz, 2023; U.S. Bureau of Labor Statistics, 2024). These statistics highlight the urgent need for innovative strategies to support persons living with cognitive impairment and their caregivers, particularly in home-based care settings that prioritize safety and quality of life.

Hine et al. (2022) found that Home Monitoring Technologies (HMTs), including artificial intelligence (AI) enabled ones, can help bridge the gap between autonomy and safety. The rapid advancements in consumer electronics and AI have introduced technologies that significantly enhance the lives of persons living with cognitive impairment. These include tools that facilitate social interactions, support telehealth services, and provide real-time tracking for safety. Technologies supporting aging in place include health monitoring devices that detect falls or changes in vital signs, sensor-based systems that track daily activity patterns, and communication tools such as voice-activated assistants (e.g., Amazon Alexa), GPS-based tracking devices to prevent wandering, interactive social robots like PARO to reduce loneliness, surveillance cameras for remote monitoring, wearable devices to track mobility and sleep, and specialized software applications that remind users to take medications or complete daily tasks. These tools have profoundly impacted the daily functioning and quality of life for people living with cognitive impairment by promoting independence, reducing caregiver burden, enabling timely medical interventions, and enhancing emotional well-being through increased connectivity and engagement (Gagnon-Roy et al., 2017; Ienca et al., 2017). Although HMTs are becoming more widely used and accepted, the adoption and implementation of HMTs introduce complex ethical challenges, including privacy violations, data breaches, loss of autonomy, and the potential dehumanization of care (Gagnon-Roy et al., 2017; Van Der Roest, 2017). Rubeis’ (2020) 4D Risk framework—focusing on depersonalization, discrimination, dehumanization, and discipline—has been used to analyze ethical issues in a range of settings, particularly in AI and algorithmic decision-making in healthcare. This framework offers a valuable lens for examining ethical concerns related to HMTs used with persons living with cognitive impairment. Discrimination emerges through structural and systemic inequalities based on attributes such as age, cognitive ability, socioeconomic status, and digital literacy. These disparities are further compounded by depersonalization, wherein individuals are perceived not as persons with unique life stories and preferences, but as abstract subjects within data-driven systems. Dehumanization similarly occurs through the datafication of individuals, reducing them to quantifiable data points and distancing them from authentic human interaction and person-centered care. This is closely tied to discipline, characterized by intense surveillance and severe violations of privacy (Rubeis, 2020). While these concepts are broadly relevant to dementia care, they are particularly concerning in the context of HMTs, as care recipients are often not fully cognizant of their implementation or practical purpose. Moreover, care partners frequently lack awareness of the technical complexities, data security risks, and privacy concerns associated with HMTs, further complicating ethical considerations. For instance, data collected by HMTs are often generalized across large datasets or used to train AI models that may not adequately represent diverse populations or the individualized nature of cognitive impairments. When such models are applied universally—without accounting for variability in age, cultural background, or disease progression—they risk misinterpretation or inappropriate responses. This lack of personalization can lead to stigma, a loss of dignity, and diminished human interaction, particularly when over-reliance on automated systems displaces meaningful, person-centered care (Rubeis, 2020). Technological interventions, such as robotic care, have the potential to improve the lives of persons living with cognitive impairment; however, they may also lead to unintended consequences, including dehumanization, reduced human interaction, and negative stigma. For instance, surveillance cameras or monitoring sensors can result in a loss of dignity and autonomy while exposing patients to privacy violations and data breaches (Rubeis, 2020). These risks highlight the importance of addressing the ethical complexities associated with HMTs. Table 1 provides an overview of the key ethical concerns of Rubeis’ 4D Risks Framework, offering operational definitions and illustrative examples drawn from the reviewed literature (see Table 1).

Table 1.

Ethical risks in HMTs for people living with cognitive impairment.

4D	Operational definition	From our findings
Dehumanization	Dehumanization in AI-based care for older adults occurs when individuals are reduced to objects, losing their personal identity and human connection. This happens when technology replaces human ­interaction, prioritizing standardized data and automation over individual experiences, leading to feelings of objectification and loss of humanity.	“Fear that electronic tracking may decrease care and social interaction, formal and informal caregivers may infer a decreased need to physically supervise” (Mangini & Wick, 2017) Assistive technologies encourage unethical trends of egoism and isolation instead of reinforcing social bonds and mutual helping, a feeling of responsibility for each other, and the search for a better-organized system of mutual care and support between different players in society. (Elger, 2019) “To some, the use of electronic surveillance is dehumanizing and threatens personal autonomy and liberty, particularly among people who cannot speak for themselves” (Landau et al., 2009) “Nevertheless, on the bad side, there is still the concern that tagging will encourage us to think of people as if they are objects, to restrict their liberty and deny them their civil rights.” (Hughes et al., 2008)
Depersonalization	Depersonalization in AI-based care for older adults is when technology, particularly AI and machine learning, reduces patients to data points, overlooking their unique experiences and needs. This results in a clinical approach that treats patients as bodies or averages, ignoring personal variations and reducing the human aspect of care.	“While they discuss ‘myths’ about telecare and thereby acknowledge the problems of depersonalization coming with ‘sensors and tags and telephone links,’ they do not consider changes in self-perception and self-image triggered by sensors close to the body. Those changes can extend the problems of depersonalization and isolation, which are particularly relevant for older adults with cognitive ­impairment or dementia.” “The use of robots in caring for persons with dementia presents ethical concerns, particularly regarding privacy and dignity. From a virtue ethics perspective, this practice risks obscuring human vulnerability and dependence, key aspects of our nature. The integration of robots into care for older adults may increase the isolation of older people, as social robots cannot fully replicate human interaction, at least in the near term.” (O’Brolcháin, 2019) “Robots cannot provide genuine care like human caregivers do, which may lead to objectification, deception, and social isolation among the elderly.” (Coin & Dubljevic, 2020)
Discrimination	Discrimination in AI-based care for older adults is when individuals are treated unfairly based on stereotypical categorizations, like ageism or minority status. It involves using data mainly from majority groups, which can overlook the unique needs of minorities or older adults, reducing them to generalized views that neglect their individual identities and preferences.	“Tele-care may only be available for those patients who lack a family caregiver or depending on the wealth of the patient. The balance between risks and benefits, as well as between beneficence and respect for autonomy in the use of assistive technologies, can only be achieved if all stakeholders are well informed.” (Elger, 2019) “The high cost of care robots and disparities in economic levels will lead to unequal access to care robot technology.” (Coin & Dubljevic, 2020) “Additionally, there are significant cultural associations in minority race and ethnic status, that combined with lower socioeconomic status and geographic locations, suggest a reduced likelihood of IAT adoption among caregivers and care recipients.” (Vollmer Dahlke & Ory, 2020) As a treatment tool, tagging and tracking technology also opens the possibility of less frequent contact with health and social care practitioners and increases the risk of even further reductions in staff in long-term care facilities. However, it is not a replacement for staff providing good quality person-centered care. (Plastow, 2006)
Disciplining	Disciplining in AI-based care for older adults refers to the use of surveillance technologies to monitor and enforce behavior, leading individuals to conform to standards set by AI. This can result in self-disciplining, where patients adapt to these expectations, often influenced by external pressures from caregivers or healthcare providers.	“Dementia patients are willing to sacrifice some privacy to use GPS to reduce the anxiety of caregivers worrying about them getting lost.” (Cooper et al., 2021) “Professionals and formal caregivers value autonomy more than family caregivers do. Family caregivers usually place less importance on autonomy compared to safety. Dementia patients in the study place a high value on their autonomy, believing that consent for GPS usage should be their own ­prerogative.” (Cooper et al., 2021) “Caregivers of people with dementia value the safety of people with dementia over preserving their autonomy, and thus they favor tracking the patient.” (Elger, 2019)

Note. AI = artificial intelligence; GPS = global positioning system; HMTs = home monitoring technologies; IAT = intelligent assistive technologies; LTC = long-term care.

Current studies examine the effects of HMTs and other assistive technologies on specific populations but often fail to consider all stakeholders involved, with some even excluding people living with cognitive impairment from the discussion entirely. Among the limited studies that focus on people living with cognitive impairment, most have primarily assessed the feasibility and effectiveness of HMTs in dementia care, particularly their role in supporting daily activities, enhancing safety, and improving quality of life for individuals with cognitive impairment. These technologies are particularly effective and feasible when implemented through a user-centered or person-centered care approach, as evidenced by the design frameworks of multiple HMT models for people living with cognitive impairment (Ienca et al., 2017). User-centered designs are frequently incorporated into HMTs to support cognition and communication challenges. For example, social robots have been developed to promote verbal engagement, while customized personal digital assistants have been designed to aid memory recall in people living with cognitive impairment (Dada et al., 2024; de Joode et al., 2010). These studies typically focus on user acceptance and clinical outcomes, a significant gap remains in understanding the ethical implications of HMTs. Existing literature on ethics in assistive technologies often focuses narrowly on specific tools, primarily GPS systems, in the context of people living with cognitive impairment (Cooper et al., 2021; Howes et al., 2022; Landau et al., 2011). While these studies provide valuable insights, they frequently lack a comprehensive perspective, failing to address broader technological advancements. They are rarely paired with modern adaptations or practical solutions to address these challenges or synthesize efforts to address ethical challenges (Mahoney et al., 2007; Portacolone et al., 2020; Vollmer- Dahlke & Ory, 2020; Wagner & Borycki, 2022). Furthermore, many reviews emphasize the perspectives of caregivers, manufacturers, or healthcare providers, while neglecting the voices and lived experiences of people living with cognitive impairment themselves (Ienca et al., 2018; O’Brolchain, 2019). Another significant gap in the literature is the lack of consideration for multiple aspects of decision-making. Most studies focus on individual stakeholders, such as family members or primary care providers, rather than adopting a collaborative decision-making framework (Ienca et al., 2018; O’Brolchain, 2019).

These limitations highlight the need for a more holistic approach to examining ethical challenges in assistive technologies, one that incorporates the perspectives of all relevant stakeholders, including people living with cognitive impairment, and evaluates solutions within the context of modern technological advancements. Our eight central themes were derived inductively during data synthesis, and the Rubeis’ 4D Risks Framework—depersonalization, discrimination, dehumanization, and disciplining risks—offered a valuable conceptual scaffold to organize and interpret the broader patterns of ethical risk. To ensure conceptual clarity, we also defined several recurring terms in our synthesis based on Rubei’s and other established frameworks: Privacy-by-design refers to the proactive integration of data protection and confidentiality measures into the technical architecture of HMTs from the outset, rather than as an afterthought (Rubeis, 2020). In dementia care, this principle underscores the importance of embedding safeguards that respect both the dignity and autonomy of people living with cognitive impairment. Dynamic consent is a model of ongoing, flexible consent that allows individuals and their care partners to make informed decisions about data use over time, reflecting changes in preferences, capacity, and context (Kaye et al., 2015). Transparency is operationalized as the clear, accessible communication of how HMTs collect, process, and share data, and how these functions shape care delivery. Transparency is central to trust-building and to reducing the informational asymmetries between technology developers, providers, and people living with cognitive impairment (Jecker, 2022). Justice and equity highlight the ethical obligation to ensure fair distribution of both the benefits and burdens of HMTs (Braveman & Gruskin, 2003). Stigma is defined as the social discrediting of people living with cognitive impairment based on assumptions of incompetence or deficit, which may be intensified when monitoring technologies depersonalize individuals or reduce opportunities for authentic human interaction (Yang et al., 2007). These concepts align closely with the risks identified in the 4D Framework. For example, discrimination emerges through systemic inequities based on age, cognitive ability, or digital literacy, which reflect broader concerns of justice and equity. Depersonalization and dehumanization occur when individuals are reduced to quantifiable data points, undermining dignity and fostering stigma. Discipline, characterized by surveillance and violations of privacy, directly underscores the need for privacy-by-design, transparency, and dynamic consent as protective mechanisms.

Therefore, under the guidance of this model, our scoping review aims to comprehensively examine the ethical considerations surrounding the adoption and implementation of HMTs for people living with cognitive impairment. This review seeks to explore how these technologies are evaluated and utilized while addressing ethical concerns from the perspectives of patients, caregivers, and manufacturers. By identifying key themes, gaps in the literature, and directions for future research, the review aims to inform the ethical development and application of HMTs grounded in person-centered care.

Method

We employed the scoping review method to synthesize information from existing studies, following the five stages outlined in Arksey and O’Malley’s framework for scoping reviews (Arksey & O’Malley, 2005). The authors followed the PRISMA-ScR checklist for reporting items in this manuscript (Tricco et al., 2018). An informal protocol, including representative search strategy (see Supplementary Appendix I, see online supplementary material) was created by the team and informally registered into Open Science Framework (OSF) (https://osf.io/tzrjx/files/osfstorage).

Stage 1: Identifying the research question

Our scoping review aims to answer the following research questions:

What ethical considerations are involved in the adoption of HMTs for people living with cognitive impairment and their caregivers? Specifically, (a) How are ethical considerations reflected in the development and adoption of HMT? (b) What are the ethical implications for both people living with cognitive impairment and their caregivers?

In addressing question (a), our goal was not to provide a technical review of development or implementation processes, but rather to examine how ethical concerns are embedded in decisions related to HMT design, adoption, and use in dementia care contexts.

Stage 2: Identifying relevant studies

Eligibility criteria

Papers meeting the following inclusion criteria were selected for the review:

1. The target population for the study is people living with cognitive impairment, including but not limited to mild cognitive impairment, AD, and other dementias and their caregivers;

2. The focus of the study is on HMTs designed to facilitate informal caregiver monitoring or tracking;

3. Study types include theoretical/conceptual work, qualitative, quantitative, mixed-method, and reviews;

4. Published in English.

Studies were excluded if they:

1. Focused on participants in long-term care or assisted living facilities; as our aim was to explore ethical issues in home-based monitoring specifically;

2. Focused primarily on technologies used for monitoring vital signs or health status (e.g., blood pressure, heart rate), as these are more commonly associated with clinical health management rather than the ethical dimensions of HMTs relevant to cognitive impairment.

Data sources

A systematic search was conducted using the following databases: PubMed, CINAHL (EBSCO), PsycINFO (EBSCO), Web of Science, ProQuest Dissertations and Theses, and the Cochrane Database of Systematic Reviews (EBSCO). The search terms included relevant terms and subject headings derived from titles, abstracts, and keywords identified in key publications and reviews related to cognitive impairment and HMTs. The search was designed by a health sciences librarian based on Peer Review of Electronic Search Strategies guidelines, then translated across all databases.

The initial database search was completed in October 2024, with an updated search conducted in February 2025 to capture newly published studies. The search covered all available dates from the inception of each database up to the present. Search results were stored in the Covidence tool for systematic reviews, and the full search strategy is publicly available in the OSF platform (https://osf.io/tzrjx/files/osfstorage/). Additionally, the references of included studies were manually reviewed to identify potentially relevant papers.

Stage 3: Study screening and selection

A total of 3,575 papers were retrieved from the database searches. After removing duplicates, a total of 2,281 papers were included for screening. The screening process followed these steps:

Step 1: Titles and abstracts were independently reviewed by two groups of reviewers (J.W., X.T., X.C., and E.H.) based on the predefined inclusion and exclusion criteria. Any disagreements were resolved by S.A. This step resulted in 244 relevant papers.

Step 2: Full-text papers were reviewed independently by the same reviewers. Disagreements were discussed and resolved with the involvement of other co-authors. Excluded full-text studies were recorded, and the reasons for exclusion were documented and reported in the scoping review. See Figure 1 for PRISMA flow diagram.

Figure 1.

PRISMA flow diagram of study selection.

Stage 4: Charting the data

Data were extracted from the full-text papers by (S.H. and A.L.) and reviewed by J.W. Any disagreements were resolved through discussion with co-authors. The extracted data included information on the technologies, target populations, study design, and ethical considerations addressed in each study.

Stage 5: Collating, summarizing, and reporting results

We categorized the included papers based on the type of monitoring technology and ethical considerations discussed. Due to the heterogeneity of outcome measures, statistical integration was not feasible. Therefore, a narrative approach was adopted to summarize the content and ethical implications of the technologies in relation to the care of people living with cognitive impairment and their caregivers (see Table 1).

Results

Overview of the included studies

A total of 110 studies were included in this scoping review, encompassing research from 30 countries. The geographical distribution revealed a significant focus on European countries, which accounted for ∼60% (66) of the studies, followed by North America with 30% (33 studies). Contributions from developing countries were comparatively limited, highlighting a notable gap in representation from these regions. Additionally, 11 studies did not specify the country of origin.

Of the 110 studies, 43 studies (39%) primarily focused on ethical considerations, explicitly stating this in their aims or research objectives. These studies explored ethical issues such as privacy, autonomy, and the use of monitoring technologies for people living with cognitive impairment. The remaining 67 studies (61%) did not primarily focus on ethics but touched upon ethical considerations briefly in their introductions or discussions.

In terms of study design, the review included a diverse array of methodologies. Among the 43 studies focusing on ethical considerations, reviews (including scoping, systematic, and narrative reviews) were the most prevalent. Approximately 52 studies across all included articles were qualitative in nature, with 28 of these specifically addressing ethics. Quantitative studies were less common, comprising 29 articles in total, of which only 7 focused on ethics. Mixed methods approach, case studies, and opinion papers were also represented, reflecting the multidisciplinary interest in ethical considerations for HMTs. Supplementary Appendix II (see online supplementary material) summarizes the included study characteristics in a table based on our data extraction spreadsheet (see Supplementary Appendix II, see online supplementary material).

The technologies explored in these studies included motion sensors, wearables, robots, and GPS systems as the most common HMTs. Among the 43 studies addressing ethics, the technologies were primarily used for surveillance, promoting independence, increasing safety, and assisting caregivers in managing activities of daily living. Non-ethics-focused studies utilized similar technologies but often emphasized general monitoring systems and enhancing quality of life through emotional and social support. The primary purposes of the HMTs varied but were mainly centered on promoting safe independence for people with dementia, assisting caregivers, and providing emotional or intellectual support. Ethical concerns were most frequently tied to the use of technologies for tracking wandering behaviors, ensuring caregiver safety, and maintaining the privacy and dignity of persons with dementia. While our review does not focus on the technical or operational aspects of how HMTs are developed, adopted, and evaluated, our findings show that ethical considerations, such as privacy-by-design, dynamic consent, transparency, equity, and stigma directly inform these processes. Similarly, the presence or absence of transparent data handling practices affects long-term trust and thus the sustained evaluation and use of these technologies.

This scoping review identified a broad range of ethical issues surrounding home-based monitoring technologies designed for people living with cognitive impairment. While prior literature has explored some of these topics, our synthesis focuses specifically on the in-home context, where technologies such as GPS, video surveillance, telepresence devices, and AI-driven systems interface closely with daily life. We identified eight central themes: privacy and data protection, autonomy and informed consent, trust and transparency, safety and freedom, stigma and dignity, justice and equity, social isolation and human interaction, and AI-specific ethical considerations. Table 1 provides an overview of the key ethical concerns of Rubeis’ 4D Risks Framework, offering operational definitions and illustrative examples drawn from the reviewed literature (see Table 1).

Privacy and data protection

Privacy emerged as a critical ethical concern across the included studies, particularly in relation to in-home monitoring via GPS tracking, video surveillance, and sensor-based technologies (Cooper et al., 2021; Ienca & Villaronga, 2019; Shore, 2021). Although some family caregivers viewed these tools as beneficial, especially for wandering prevention (Ienca & Villaronga, 2019; White & Montgomery, 2014), others noted their intrusive nature and potential to erode informational, physical, and attentional privacy. Studies also highlighted how continuous monitoring can transform the home—traditionally a personal sanctuary—into an environment of perpetual surveillance (Shore, 2021). Multiple articles (Ienca & Villaronga, 2019; Simon et al., 2022) cautioned that if devices lack rigorous security safeguards, personal data could be hacked, stolen, or even used for unauthorized surveillance. Legitimate reasons for in-home monitoring (e.g., tracking someone’s health status) risk overreach if “bycatching” collects information about family members or visitors who never consented. Authors recommended “privacy-by-design” approaches to minimize data storage, limit access to designated personnel, and maintain transparency about any third-party sharing (Ienca et al., 2016; Ienca & Villaronga, 2019). Privacy-by-design strategies, such as limiting data storage to real-time monitoring or restricting access to essential users, were proposed to mitigate these risks (Elger, 2019; Ienca & Villaronga, 2019). Nonetheless, concerns remained about stigmatization and dignity. Even well-intentioned “protective” practices (e.g., Project Lifesaver) can inadvertently label people living with cognitive impairment as perpetually “vulnerable” (Shore, 2021), thereby fueling paternalistic or stigmatizing views (Shore, 2021). Furthermore, cybersecurity vulnerabilities, such as the potential for data breaches, hacking, and external control were raised, emphasizing that if certain monitoring devices or robots lack robust security features, personal information could be compromised (Ienca & Villaronga, 2019).

A few articles (Ienca & Villaronga, 2019; Simon et al., 2022) cautioned that if devices lack rigorous security safeguards, personal data could be hacked, stolen, or even used for unauthorized surveillance. Legitimate reasons for in-home monitoring (e.g., tracking someone’s health status) risk overreach if “bycatching” collects information about family members or visitors who never consented. This theme aligns with the Disciplining aspect of the 4D model, as it causes individuals to conform to the standards set by these surveillance technologies and sacrifice their privacy in order to appeal to caregivers (Rubeis, 2020).

Importantly, caregiver perspectives often prioritized safety and security over privacy, while people living with cognitive impairment perspectives (when elicited) emphasized the intrusive and dignity-eroding aspects of surveillance. Professional stakeholders, by contrast, tended to advocate for privacy protections as a safeguard of ethical standards. This divergence underscores a persistent tension between stakeholders that the literature describes but rarely resolves. Moreover, most evidence was descriptive and caregiver-centered, with limited direct accounts from people living with cognitive impairment, suggesting gaps in representativeness and strength of evidence. In this context, privacy-by-design operates both as an internal factor (embedded technical safeguards) and an external factor (shaped by caregiver practices, professional norms, and regulatory requirements). This dual role highlights its potential to bridge stakeholder concerns but also exposes limitations when applied inconsistently.

Autonomy and informed consent

Ensuring autonomy often conflicted with the goal of safety in home settings, where caregivers must independently balance user independence with the risk of wandering (Cooper et al., 2021; Landau et al., 2011). Several studies emphasized dynamic consent models that involve people living with cognitive impairment in decisions around adopting or discontinuing HMTs (Elger, 2019; Landau & Werner, 2012). Family caregivers, however, tended to prioritize safety, sometimes resorting to covert or paternalistic practices (Cooper et al., 2021). In contrast, professionals generally placed more value on respecting individual autonomy (Landau et al., 2009; White & Montgomery, 2014). Articles called attention to the ethical complexity of situations wherein cognitive decline reduces the capacity for meaningful consent, leading to either proxy decision-making or, in some instances, unwitting acceptance of surveillance. Sources also pointed out that, in situations where an individual’s physical or cognitive condition has significantly deteriorated, fully informed consent is difficult to obtain (Mahoney et al., 2007; Nagel & Remmers, 2012). This underscores how “hard paternalism” (coercing someone still capable of choice) differs from “soft paternalism” (acting when capacity is substantially impaired) (Elger, 2019). In some publications, timing was emphasized: adopting monitoring solutions earlier, when the individual can still articulate preferences, allows for a more ethically grounded approach (Landau et al., 2011; Landau & Werner, 2012). This central theme is linked to the Disciplining and Dehumanization aspects of the 4D model, as caregivers sometimes prioritize safety over autonomy. The overuse of these monitoring technologies can threaten autonomy, especially for individuals who cannot speak for themselves (Rubeis, 2020).

Overall, caregiver perspectives were more likely to accept paternalistic practices to reduce perceived risk, while people living with cognitive impairment and professional perspectives placed stronger emphasis on autonomy. This divergence points to an unresolved debate: whether safety justifies infringements on autonomy, or whether autonomy should remain the higher priority. The literature often described this tension but offered few concrete strategies for resolution, highlighting a need for more robust, comparative studies rather than primarily descriptive accounts. Once again, the evidence base was dominated by caregiver perspectives, limiting insights into people living with cognitive impairment’s own experiences.

Trust and transparency

Trust was frequently identified as crucial for encouraging acceptance and sustained use of in-HMTs (Coin & Dubljevic, 2020; Ienca & Villaronga, 2019). Studies underscored the necessity of transparent communication regarding data collection, storage, and potential third-party sharing (Shore, 2021). Where technologies employed anthropomorphic designs—such as pet-like or humanoid robots—users sometimes misread empathetic features as genuine care, raising questions about deception (Coin & Dubljevic, 2020). Continuous, adaptive consent mechanisms were proposed to help maintain trust over time (Berridge et al., 2022; Elger, 2019). However, the complexity of AI-driven or smart home platforms often left families uncertain about system decisions, undermining confidence in how and why certain data were captured and analyzed. This theme closely aligns with the Dehumanization aspect of the 4D model, as these technologies can be deceiving and reduce social interaction for people living with cognitive impairment (Rubeis 2020).

The literature suggests that trust is shaped differently for stakeholders: caregivers valued transparency primarily to reassure themselves of safety, while people living with cognitive impairment were more sensitive to how transparency affects their dignity and autonomy. Developers and providers, meanwhile, framed transparency as a technical or regulatory requirement rather than a relational necessity. These differing emphases create gaps in expectations that can undermine trust. The strength of evidence here remains modest, with most accounts based on small qualitative studies rather than systematic evaluations of trust-building interventions. people living with cognitive impairment voices were rarely included directly, leaving caregiver views as the primary lens.

Safety and freedom

A recurring theme was the tension between maximizing safety and preserving personal freedom. While recognizing the importance of “least restrictive” interventions, some references emphasized that even these can unintentionally curtail freedom if caregivers rely too heavily on digital tracking rather than periodic check-ins (Landau et al., 2010; Welsh et al. 2003). GPS tracking was widely seen as a comparatively less restrictive solution for preventing or managing wandering behaviors (Hughes et al., 2008; Landau et al., 2010). A few sources (20, Cooper et al., 2021; Elger, 2019) noted that “wandering within limits” can be beneficial, providing exercise and a degree of autonomy that fosters well-being. Consequently, any decision to impose continuous surveillance should weigh the positive aspects of modest risk-taking against the potential for over-monitoring. Debates revolved around beneficence and non-maleficence, wherein the perceived benefits of minimizing harm had to be balanced against possible infringements on independence (Hughes et al., 2008; Mulvenna et al., 2017). Caregivers reported reduced anxiety and an enhanced sense of security when using these tools (Cooper et al., 2021; Landau et al., 2010). Yet, some researchers warned that such reliance might institutionalize paternalistic care practices or subtly confine individuals to monitored spaces (Landau et al., 2010; Welsh et al., 2003). This theme relates to the Dehumanization and Disciplining aspects of the 4D model, because HMTs can limit the freedom of people living with cognitive impairment in favor of safety by monitoring and enforcing behavior (Rubeis, 2020).

Caregivers frequently perceived safety technologies as reducing their burden and anxiety, while people living with cognitive impairment often experienced the same technologies as restrictive or infantilizing. The literature highlights this contradiction but provides little empirical evidence of how to balance these competing priorities. Most studies offered descriptive accounts without rigorous evaluation of whether “least restrictive” approaches effectively preserve both safety and freedom.

Stigma and dignity

Stigmatization was often linked to visibly identifying technologies like GPS bracelets or robots, which risk portraying individuals with dementia as akin to “lost children” (Cooper et al., 2021; Elger, 2019). Several studies recommended discreet, user-centered design to avoid exacerbating negative public perceptions (Shore, 2021; Yang & Kels, 2017). Similarly, in-home surveillance cameras could undermine dignity by turning personal spaces and behaviors into objects of observation (Mulvenna et al., 2017). Overall, the literature stressed the importance of design approaches and deployment practices that reduce labeling and intrusion, ensuring that people with dementia retain a sense of self-worth and personal agency. This theme aligns with the Dehumanization aspect of the 4D model, as HMTs can disregard individual experiences, leading to feelings of objectification and loss of humanity.

Stakeholder perspectives diverged here as well: caregivers often saw technologies such as GPS bracelets as neutral safety tools, while people living with cognitive impairment and advocacy groups highlighted their stigmatizing symbolism. Evidence was mixed, with some studies reporting enhanced feelings of security, while others noted reduced dignity. The descriptive nature of this evidence makes it difficult to determine under what conditions stigma predominates over perceived benefit.

Justice and equity

Issues of equitable access surfaced repeatedly, reflecting how cost, broadband availability, and digital literacy impact the implementation and utility of HMTs (Elger, 2019; Vollmer Dahlke & Ory, 2020). Rural or low-income households often lack reliable internet infrastructure, limiting the feasibility of remote sensors or AI-driven devices. Cultural attitudes toward surveillance can further complicate adoption, with acceptance varying widely according to social norms and family structures (Vollmer Dahlke & Ory, 2020). Several articles advocated for affordability measures and the inclusion of culturally responsive designs that do not widen existing healthcare disparities (Elger, 2019). This theme is closely linked to the Discrimination aspect of the 4D model, because certain HMTs may not be accessible for all people living with cognitive impairment, and they may be developed in a way that overlooks the needs of minorities or neglects individual identities and preferences (Rubeis, 2020).

Justice and equity concerns were primarily raised at a systems level, such as infrastructure, affordability, literacy. Caregivers often noted access barriers, but people living with cognitive impairment voices were rarely included directly. The literature stressed distributive justice but offered limited concrete strategies beyond affordability and broadband expansion. This suggests a strong recognition of inequity but weak evidence on effective remedies.

Social isolation and human interaction

While many assistive technologies support independence, they could inadvertently contribute to decreased human interaction (Astell & Semple, 2019; Berridge et al., 2022). The ease of remote checks or robotic support sometimes reduced face-to-face visits, thereby limiting social engagement (Mulvenna et al., 2017; Robillard et al., 2020). In the home context, this phenomenon can be amplified as families or friends perceive less urgency to visit, believing the individual is “safe” under constant monitoring. Researchers cautioned that technologies should complement—rather than replace—meaningful human contact, lest people living with cognitive impairment become more isolated and disengaged from their communities. This theme aligns with the Dehumanization and Depersonalization aspects of the 4D model, because HMTs can lead to human interaction being replaced by technology, which reduces personalization to the individual. The lack of social connection and generalized experiences leads to objectification and unmet needs (Rubeis, 2020).

Caregivers often reported reassurance and reduced stress from remote monitoring, while people living with cognitive impairment described feelings of neglect or reduced visits. This contradiction illustrates how technologies may simultaneously reduce caregiver burden and increase isolation for people living with cognitive impairment. Evidence was again descriptive, largely based on interviews or case studies, with no longitudinal data to assess sustained impacts on social connectedness.

AI-specific ethical considerations

AI-driven systems introduced additional complexities for home monitoring. Certain devices, such as care robots and wearable analytics tools, potentially override user preferences to ensure safety (Hoorn, 2015; Ienca & Villaronga, 2019). Animal-like or humanoid interfaces can lead users to assume the device displays empathy, raising ethical debates about authenticity and/or deception (Coin & Dubljevic, 2020; Portacolone et al., 2020). Trust was further strained if algorithms’ decision-making processes were not fully transparent (Berridge et al., 2022; Hoorn, 2015). Equitable access presented a barrier, with broadband limitations and hardware costs precluding adoption in underserved areas (Vollmer Dahlke & Ory, 2020). Critiques extended to the potential for social isolation, as AI could substitute in-person caregiving and erode direct interpersonal connections (Astell & Semple, 2019; Robillard et al., 2020). Lastly, some articles emphasized that there are currently no robust guidelines for overseeing AI in dementia care, leaving room for biases in training datasets and opaque “black box” algorithms (Portacolone et al., 2020; Mulvenna et al., 2017).

AI-specific concerns were discussed mainly in theoretical or speculative terms rather than based on empirical evaluations. This reflects both the novelty of AI applications in dementia care and a lack of robust evidence to assess risks such as bias, opacity, and authenticity. The literature therefore flags significant ethical risks but does not yet provide evidence on effective mitigations, leaving tensions between innovation and protection unresolved.

Discussion

This scoping review included 110 studies from 30 countries, some of which primarily focused on ethical considerations, while others touched upon these issues in the context of home-based monitoring technologies for individuals with cognitive impairments. Ethical concerns such as privacy, autonomy, and informed consent were central across the included studies. Many studies emphasized the balance between ensuring safety through technology and preserving individual autonomy, as well as the challenge of maintaining trust and transparency in systems that could be perceived as intrusive or paternalistic. Additional themes included concerns about stigma and dignity, especially related to visible monitoring tools, and the potential for social isolation due to decreased human interaction. Inequities in access to technology were also noted, with rural and low-income populations facing barriers. Finally, AI-driven systems raised a unique ethical dilemma regarding decision-making, data privacy, and the replacement of human caregiving with robotic solutions.

Privacy concerns present complex ethical dilemmas. While many dementia patients are willing to sacrifice some privacy for practical benefits, such as GPS tracking to alleviate caregiver anxiety, there is a clear distinction between familial care and surveillance by external entities. It should also be included that caregivers tend to prioritize safety over privacy but must navigate the tension between maintaining independence and protecting dignity. Privacy breaches, though inherent in surveillance, can be justified when the intent is to support autonomy, such as allowing individuals to live at home longer (Clark et al., 2023). When implementing technology, it is important to limit data access to those who need it and to ensure that security is maintained. Violations of privacy may undermine a person’s sense of agency, dignity, and independence, leading to risks of caregiver overreach, anxiety, and feelings of infantilization, which can ultimately impact trust (van der Geugten & Goossensen, 2020). Furthermore, maintaining confidentiality and ensuring compliance with regulations such as HIPAA when technology is used to share health data updates with providers, as well as restricting unnecessary data sharing are essential to safeguarding both privacy and dignity in this context.

A central ethical concern for the utilization of HMTs for adults with dementia is the appropriate balance between autonomy and safety. HMTs may empower individuals by supporting independent living within reason and aging in place. However, dementia-related cognitive decline may impair a person’s ability to make informed choices, raising safety concerns. Multiple included studies report that caregivers often prioritize safety over personal freedom. In some cases, it may lead to caregivers adopting paternalistic practices. Some believe that ensuring safety through surveillance and monitoring may restrict certain personal freedoms for people living with cognitive impairment. The ethical dilemma arises in determining to what degree HMTs should intervene, with respect to protecting individuals from harm while preserving independence and dignity. This balance must be negotiated carefully, while considering the potential risks of both overprotection and reasonable risk taking (Clark et al., 2023).

While assistive technologies promise greater independence, they may unintentionally contribute to reduced social engagement (Astell & Semple, 2019; Berridge et al., 2022). Our findings indicate that caregivers perceive AI-driven monitoring tools as a “proxy presence,” which can lead to fewer in-person check-ins. Rubeis (2020) warns against an over-reliance on AI replacing human caregiving roles, as AI-based automation shifts care from a relational to a transactional model. The ethical risk here is that AI interventions might be used as a cost-saving mechanism, reducing the presence of trained caregivers in favor of automated tracking and remote monitoring. In-person connection remains critical for emotional well-being and cognitive health. However, it’s important to recognize that, in certain contexts—such as during the COVID-19 pandemic—technology can bridge the gap when physical visits are not possible. Virtual platforms enabled individuals to maintain social connections and access care remotely when traditional interactions were restricted (Barbosa et al., 2024). While there is a real risk of isolation, especially for persons living with cognitive impairment, well-designed technology interventions can alleviate some of the burden on informal caregivers who are already operating under conditions of high burnout, anxiety, and unmet care needs. National data consistently highlight these unmet needs and the psychosocial distress experienced by caregivers, underscoring the importance of viewing technology not as a replacement for in-person care, but as a flexible support tool that can enhance care continuity and reduce caregiver strain when thoughtfully implemented.

Our study highlights equity issues related to technology access. Factors such as cost of technology, digital literacy disparities, and broadband limitations disproportionately affect marginalized groups (Elger, 2019; Vollmer Dahlke & Ory, 2020). This aligns with Rubeis’ argument that AI in care for older adults is often trained on majority-group data, leading to bias in predictive analytics and care recommendations. Moreover, cultural attitudes toward surveillance vary, with some communities perceiving constant monitoring as invasive rather than protective (Vollmer Dahlke & Ory, 2020). If these cultural nuances are not embedded into AI training datasets, certain populations may be either over-monitored or underserved by emerging technologies. To address these disparities, it is crucial to prioritize the development of affordable, culturally sensitive technologies that are accessible regardless of geographic location or socioeconomic status (Vollmer Dahlke & Ory, 2020). Efforts to reduce these disparities should focus not only on making technology more affordable but also on enhancing digital literacy and infrastructure to ensure no further marginalization of already vulnerable populations.

Obtaining informed consent for the implementation of HMTs requires a nuanced, person-centered approach, recognizing that consent is a continuous and dynamic process. It is essential to assess an individual’s capacity for informed decision-making, considering cognitive function and varying levels of understanding. Consent should not be seen as a one-time event but rather an ongoing conversation that adapts to the person’s changing needs and cognitive abilities. As care needs evolve, the level and type of technological support may also need to change—potentially requiring renewed consent. However, these transitions often occur when a person’s capacity for understanding and decision-making may be compromised, coinciding with heightened caregiver burden. This creates a potential ethical tension or “tug of war” between respecting the rights and autonomy of the individual and addressing the practical needs and responsibilities of caregivers and care providers. Additionally, when incorporating AI-driven systems, there are unique ethical considerations, particularly the potential for deception and depersonalization. AI technologies often operate with a level of autonomy and may unintentionally influence decision-making, raising concerns about the individual needs and preferences of people living with cognitive impairment being subsumed under algorithmic norms. While AI allows for personalized adjustments, its reliance on large datasets from general populations may fail to capture the unique progression of cognitive impairment in diverse individuals. If caregivers prioritize AI-generated recommendations over lived experiences, the ability to adapt interventions dynamically to individual variability in dementia trajectories may be reduced. Also, the risk of reliance on technology that lacks human empathy or the ability to recognize cognitive decline emphasizes the need for careful oversight and collaboration to ensure that an individual can maintain autonomy and dignity.

A critical concern in our review is the implicit disciplining effect of continuous surveillance. Rubeis draws on Foucault’s concept of the panopticon, suggesting that elderly individuals adapt their behaviors to conform to monitoring expectations. Our findings align with this notion—families often implement AI-based tracking to reduce anxiety about wandering, but at the cost of limiting spontaneous activity (Landau et al., 2010; Welsh et al. 2003). Moreover, people living with cognitive impairment may internalize surveillance, leading to self-restriction of movement and behaviors perceived as “risky” (Elger, 2019; Lodha & De Sousa, 2020). The challenge is to balance safety without eroding autonomy, ensuring that monitoring technologies support rather than confine individuals.

This study has several significant strengths that enhance its quality and reliability. The research question is clearly defined, focusing specifically on the ethical considerations of adopting HMTs for people living with cognitive impairment and their caregivers. This clear focus ensures that the review stays on track, addressing the most important issues. Another strength is in the use of multiple databases, which broadens the scope of included studies and reduces the risk of selection bias. The two-step screening process conducted further strengthens the study’s methodology by providing an additional layer of examination, increasing the reliability for study selection. Additionally, the data extraction process was thoroughly outlined, with multiple reviewers involved to ensure consistency and validity, and any disagreements were resolved through discussion, reinforcing the study’s methodological thoroughness. Also, the study considers both caregiver and individual perspectives, capturing a holistic view of the issue at hand. The inclusion of a wide variety of study types ranging from theoretical and conceptual work to qualitative, quantitative, and mixed-method studies broadens the scope of insights and helps paint a more comprehensive picture of the available research. The diverse literature sources included articles, book chapters, and empirical studies, which further enrich the study’s findings and contribute to a well-rounded analysis. These strengths collectively ensure that the study provides a robust exploration of ethical considerations in caregiving technologies.

The limitations of this review primarily stem from several key factors. First, by focusing exclusively on studies published in English, the review may have overlooked relevant research conducted in other languages, potentially limiting the scope of findings. However, more than half of the studies did include non-US populations, increasing our confidence in the extracted themes. The exclusion of studies involving older adults residing in assisted living facilities restricts the applicability of results to this important subset of the population. This review also omitted studies centered on monitoring vital signs or health status which is justifiable but may inadvertently exclude technologies that are highly pertinent to the broader caregiving context. Additionally, the variability in study designs and populations presents challenges in drawing generalized conclusions, as differences in methodology and sample characteristics can limit the comparability of results. Finally, ethical considerations associated with different technologies may differ based on factors like type, context, and geographic location, which make it difficult to summarize these concerns effectively without generalization or simplification.

Some of the studies included in this review were theoretical in nature and did not always employ systematic methods or empirical data to assess ethical issues. This introduces potential bias related to author perspective, disciplinary lens, and interpretation of ethical frameworks. Additionally, a majority of studies originated from high-income countries—particularly in Europe and North America—where cultural norms, digital infrastructure, and privacy regulations may differ substantially from those in underrepresented regions. This geographic skew could limit the generalizability of some ethical themes to global or resource-limited settings.

Moreover, among the studies not explicitly focused on ethics, ethical issues were often discussed briefly or anecdotally in discussion sections, which may result in underreporting or uneven emphasis on certain themes. While our review synthesizes key ethical concerns based on available evidence, we caution that these findings should be interpreted within the context of these limitations.

Future discussions should address how best to address decision-making concerns to ensure that HMT use remains person-centered. Further research is needed to prioritize this approach, particularly as HMTs continue to evolve. Collaborating with technology developers is crucial to align ethical frameworks with the design of HMTs for people living with cognitive impairment. The complexities of ethical considerations in caregiving technologies also highlight the need for more detailed research, as a narrative synthesis may overlook subtle issues best explored through quantitative or comparative methods. Given the rapid pace of technological advancements, it is crucial to regularly update the literature search to capture emerging trends. Additionally, categorizing technologies more granularly, such as by tracking, behavioral monitoring, and safety functions, would help clarify which ethical concerns are linked to specific technological categories and provide a clearer understanding of the ethical landscape. Building on these observations, our review points to three key areas for actionable guidance. First, future ethics research must adopt methodological approaches that directly include people living with cognitive impairment, using strategies such as adapted consent procedures, participatory or co-design methods, and triangulation across stakeholder groups. This would help rebalance a literature base that currently privileges caregiver perspectives and ensure that autonomy, dignity, and lived experience remain central. Second, design principles for HMTs should prioritize cultural responsiveness and accessibility, including attention to affordability, broadband access in rural or underserved areas, and stigma-reducing design features. Developers, caregivers, and people living with cognitive impairment should collaborate early in the design process to embed privacy-by-design, dynamic consent, and transparency into the architecture of technologies rather than treating them as afterthoughts. Third, policy priorities may include stronger data governance and consent frameworks tailored to dementia care contexts. This entails clear accountability for how data are stored, shared, and used; regulatory oversight to prevent exploitation or misuse; and policies that acknowledge consent as a dynamic, ongoing process rather than a one-time transaction. Together, these steps would not only fill gaps identified in the literature but also provide a roadmap for ethically grounded research, practice, and policy.

Conclusion

HMTs show tremendous potential to preserve autonomy, support aging in place, increase safety for people living with cognitive impairment, and reduce caregiver burden. However, their successful integration into dementia care requires careful attention to the ethical complexities that arise at the intersection of technology design, user needs, and caregiving contexts. The dynamic interplay between caregiver and care recipient preferences, varying levels of cognitive capacity, risks of depersonalization and dehumanization, and the potential erosion of human connection must be critically examined. To fully realize the promise of HMTs, future efforts must prioritize inclusive design, ethical safeguards, and ongoing dialogue among stakeholders—including people living with cognitive impairment, caregivers, clinicians, and technologists—to ensure that technological innovation supports, rather than undermines, person-centered and equitable dementia care.

Supplementary material

Supplementary data are available at The Gerontologist online.

Funding

1. Fostering Ethical Adoption of Artificial Intelligence-Enabled Assistive Robots (AIAR) Grounded in Person-Centered Dementia Care (R21EB036448).

2. Effectiveness and adoption of a Smart home-based social assistive robot for care of individuals with AD (R01AG075892).

Conflict of interest

None declared.

Data Availability

All materials needed to replicate this review, including the informal protocol, inclusion/exclusion criteria, and data-extraction template are available on the OSF project page: https://osf.io/tzrjx/files/osfstorage. Representative search strategies are provided in Supplementary Appendix I (see online supplementary material) and mirrored on OSF. Full-text PDFs of included articles are third-party copyrighted and cannot be redistributed; complete citations are provided in the manuscript. The team developed an informal, a priori protocol and posted it to OSF prior to data extraction (see link above). PROSPERO registration is not required for a scoping review.