*2.2.2 Acceptance models adapted for social robots*

*Collaborative and Humanoid Robots*

how it can be useful to them.

computer screen [22].

*2.1.4 What is different about social robots?*

research practice will now be presented.

*2.2.1 Traditional technology models of acceptability*

**2.2 Models used to explain and examine robot acceptability**

by emotional reactions rather than rational decisions [24].

regarding its capacity and its level of diffusion in society [21]. The relatively little diffusion of robots in society currently impacts acceptability factors such as attitudinal beliefs about the technology. Lack of familiarity also impacts how at ease people are with a technology and their need for information in order to understand

The acceptance of robots is also impacted by the additional features that robots possess in comparison to traditional technologies. Robots are embodied devices that share the space with their users. Embodiment impacts how robots are perceived, and how people interact with robots and the type of relationship humans can build with them. OA prefer to interact with an embodied social robot rather than a

Robots are also designed to behave and move more autonomously than traditional technologies. In addition, robots are designed with the intention to promote social interaction between themselves and the human user. These features impact their acceptability because the robot has a social presence and to be effective it must cause the human user to perceive it as a social identity. Variables including social presence and perceived sociability will be discussed further below in relation to the theoretical models that have used these concepts to explain and predict technology acceptance. Some examples of these models and how they have been used in

The Theory of Planned Behaviour was developed from the Theory of Reasoned Action [23]. This model proposes that people are influenced to behave in a certain way by making rational decisions about the personal and social outcomes that they anticipate as a result of their behaviour. This model may be helpful to explain the intention to use robots [4]. However, it cannot adequately explain the acceptance of robots, without being adapted. Because the decision to use robots can be impacted

The Technology Acceptance Model (TAM) [25] was derived to focus on the acceptance of computerised information systems in workplace contexts. TAM has been reported to explain 40% of variance in acceptance [26]. This model regards acceptance in terms of Intention to Use (ITU) the technology. Actual usage of the technology may or may not follow a potential user having ITU a robot. ITU is considered to be dependent on the user's attitudes towards the technology which is derived from their assessment of its Perceived Usefulness (PU) and on users' Perceived Ease of Use (PEOU). TAM was developed further [27] to include the impact of social influence, facilitating conditions, and habitual usage on technology acceptance. The Unified Theory of Acceptance of Technology (UTAUT) offers a social psychological approach that can explain 70% of acceptance variance [27]. It uses constructs from eight previous theoretical models. It has four independent variables: performance expectancy, effort expectancy, social influence, and facilitating conditions. The independent variables affect the dependent variables: ITU and actual usage. Their effect on ITU and actual usage is moderated by gender, age, experience, and voluntariness of use. The UTAUT has been criticised as not being parsimonious and for combining highly correlated variables that provide an artifi-

**20**

cially high variance [4].

The Almere Model was developed using experiments (n = 4) that involved three robots, to test acceptance of social robots by OA (188) [19]. The Almere model extends and adapts the UTAUT. It has eleven constructs which are defined in **Table 2**. These constructs enable the measurement of acceptability to focus on aspects of the technology pertinent to social robots and the perception of humans towards them as autonomous embodied social entities. Heerink et al. [19] found that the model was strongly supported accounting for 59–79% of the variance in usage intentions and 49–59% of the variance in actual use. However, this was achieved using equation modelling on four separate databases without confirming their similarities [4]. It has also been suggested that users' beliefs about their self-efficacy to use and control the robot are underrepresented in the Almere model [18].

The Almere model has been used to explore levels of acceptance and engagement with a humanoid robot, Matilda [7]. This study involved people with dementia (n = 115) living in care homes (n = 4). It used a mixed-method longitudinal experience in which the reactions of participants were coded (n = 8304) according to emotional visual behavioural and verbal engagement measures. Participants used Matilda in three repeated 4–6 hour stages of field trials and as a result of the feedback received from participants, Matilda was further developed and improved.

The Model of Social Robot Acceptance was developed by expanding and adapting the theory of planned behaviour using some elements of the UTAUT and including factors relevant to robots [4]. This model utilises eight constructs that are summarised in **Table 3**. De Graaf's model regards users' attitudes as being comprised of both hedonic and utilitarian beliefs and it includes consideration of control beliefs and that people may use a robot, and any other assistive devise, due to habitual behaviour rather than making a rational decision to use it. This


#### **Table 2.**

*Constructs in the Almere Model (Adapted from Heerink et al. [29]).*


#### **Table 3.**

*Constructs in the model of social robot acceptance [4].*

model was tested by de Graaf [4] with respondents from an online questionnaire (n = 1248) of whom, 24.7% were aged over 60. This found that intention to use the robot increased when people believe: they have the skills to use the robot; that using it would be enjoyable; it would increase their status; when the robot would be less sociable; and cause less worry about privacy. However, only self-efficacy significantly impacted user intention to use the robot. De Graaf [4] concludes that the model is a useful guide to the identification of acceptance variables but that it needs further development and testing with other data sets. The Model of Social Robot Acceptance was used by Korblett [18] in the study examining telepresence robots that has been described above.

One limitation of applying the Almere model and the Model of Social Robot Acceptance to OA and people living with dementia is that they do not take account of disability and the fact that OA and people with dementia may have additional needs concerning technology usage. It has been argued that everyday and psychosocial functioning are better predictors of robot acceptance than chronological age per se [16]. The usage of technology is impacted by a person's level of cognitive ability which affects their psychomotor speed, domain knowledge, and visual memory [11]. People with dementia may also experience declining touch sensitivity and less ability to execute accurate and discrete movements [6, 30]. Learning new technologies can also be difficult with symptoms of dementia such as reduced working memory, information processing ability and speed, and a lack of ability to disregard unwanted information [31]. Therefore, traditional models of technology acceptability and robot-specific models may be inadequate to explain the acceptance factors pertaining to some OA, particularly those with dementia.

Models of Gerontechnology acceptance have been developed which do consider physical and cognitive health. Gerontechnology has been defined as electronic or digital products or services that can increase independent living and the social participation of OA [30]. Two examples of models are described below.

#### *2.2.3 Acceptance models adapted for gerontechnology*

The Senior Technology Acceptance Model (STAM) [30] was developed with data collected by personal interviews with community-dwelling people (n = 1012)

**23**

nological models.

*Examining Social Robot Acceptability for Older Adults and People with Dementia*

over 55 years living in Hong Kong. STAM extends the constructs of previous technology acceptance models by adding age-related health and ability constructs: self-reported health conditions, cognitive abilities, attitudes towards ageing and life satisfaction, social relationships, and physical functioning. Chen and Chan [30] found that STAM can explain 68% of variance in gerontechnology usage and that facilitating conditions directly impacted the usage behaviour of participants. This suggested that the support of other people, knowledge, and guidance, directly impacts the gerontechnology usage of OA. The STAM has been used to appraise changes in technology acceptance in a randomised controlled trial that involved people with dementia (n = 103) living in residential care facilities (n = 7) [12]. The link between factors impacting the acceptance of a social robot, Kabochan, was assessed to ascertain if participant's attitudes and beliefs towards the robot would be impacted by the amount and the quality of engagement they had with the robot. They found that PEOU changed but beliefs and attitudes remained unchanged

The Model of Gerontechnology acceptance [32] encompasses both disability and aspects of a person's living environment. It was developed within the discipline of social gerontology, developed based on a study involving OA (n = 67) living in their own homes who were interviewed in-depth to ascertain their usage of and experience with a range of assistive technology (AT). AT was defined as being any device or system that allows an individual to perform a task that they would otherwise be unable to do. McCreadie and Tinker [32] argue that acceptability of AT is impacted by the interaction between a 'felt need' for assistance, and recognition of the product's quality regarding its efficiency, reliability, simplicity and safety, availability, and cost. They found four user attributes that were particularly relevant: OA disability, living arrangements, carer needs, and personal motivations and preferences. These attributes combined to generate a felt need for assistance from the technology, which combined with external factors that impacted access to the AT: information, contacts with suppliers, knowing what help is available. McCreadie and Tinker [32] emphasis that for AT to be acceptable, an OA must have information about how an AT might address their needs and be beneficial to them. In addition, OA acted in reciprocal relationships, with a strong desire for independence and to exercise their autonomy and agency in deciding whether to accept and AT [32]. The importance of autonomy and the agency of OA and people with

It is apparent that although gerontechnology models of acceptance do not address constructs that are particular to social robots as embodied social presences, they do highlight other factors that are pertinent to the way OA and people with dementia evaluate the acceptance of technology. Indeed, all the types of conceptual models currently available to guide empirical research in this context appear to have merits and limitations. It is also interesting to note that a recent systematic review, that included n = 74 studies, found that most studies that examine the technology acceptance, adoption, and the usability of information communication technology for people with dementia and their caregiver partners, do not report being underpinned by theoretical models and most use bespoke approaches to measuring acceptability [20]. Therefore, it is evident that a more consistent approach to examining [20] and conceptualizing robot acceptability for OA and people with dementia needs to be developed. It may be helpful for this approach to include constructs that are pertinent to robot usage and includes factors used in gerontech-

The empirical research will now be examined to identify the factors that impact robot acceptability and what is important to OA and people with dementia and is

likely to impact their decision whether to use a robot.

*DOI: http://dx.doi.org/10.5772/intechopen.98365*

despite engagement with the robot.

dementia is discussed further below.

#### *Examining Social Robot Acceptability for Older Adults and People with Dementia DOI: http://dx.doi.org/10.5772/intechopen.98365*

over 55 years living in Hong Kong. STAM extends the constructs of previous technology acceptance models by adding age-related health and ability constructs: self-reported health conditions, cognitive abilities, attitudes towards ageing and life satisfaction, social relationships, and physical functioning. Chen and Chan [30] found that STAM can explain 68% of variance in gerontechnology usage and that facilitating conditions directly impacted the usage behaviour of participants. This suggested that the support of other people, knowledge, and guidance, directly impacts the gerontechnology usage of OA. The STAM has been used to appraise changes in technology acceptance in a randomised controlled trial that involved people with dementia (n = 103) living in residential care facilities (n = 7) [12]. The link between factors impacting the acceptance of a social robot, Kabochan, was assessed to ascertain if participant's attitudes and beliefs towards the robot would be impacted by the amount and the quality of engagement they had with the robot. They found that PEOU changed but beliefs and attitudes remained unchanged despite engagement with the robot.

The Model of Gerontechnology acceptance [32] encompasses both disability and aspects of a person's living environment. It was developed within the discipline of social gerontology, developed based on a study involving OA (n = 67) living in their own homes who were interviewed in-depth to ascertain their usage of and experience with a range of assistive technology (AT). AT was defined as being any device or system that allows an individual to perform a task that they would otherwise be unable to do. McCreadie and Tinker [32] argue that acceptability of AT is impacted by the interaction between a 'felt need' for assistance, and recognition of the product's quality regarding its efficiency, reliability, simplicity and safety, availability, and cost. They found four user attributes that were particularly relevant: OA disability, living arrangements, carer needs, and personal motivations and preferences. These attributes combined to generate a felt need for assistance from the technology, which combined with external factors that impacted access to the AT: information, contacts with suppliers, knowing what help is available. McCreadie and Tinker [32] emphasis that for AT to be acceptable, an OA must have information about how an AT might address their needs and be beneficial to them. In addition, OA acted in reciprocal relationships, with a strong desire for independence and to exercise their autonomy and agency in deciding whether to accept and AT [32]. The importance of autonomy and the agency of OA and people with dementia is discussed further below.

It is apparent that although gerontechnology models of acceptance do not address constructs that are particular to social robots as embodied social presences, they do highlight other factors that are pertinent to the way OA and people with dementia evaluate the acceptance of technology. Indeed, all the types of conceptual models currently available to guide empirical research in this context appear to have merits and limitations. It is also interesting to note that a recent systematic review, that included n = 74 studies, found that most studies that examine the technology acceptance, adoption, and the usability of information communication technology for people with dementia and their caregiver partners, do not report being underpinned by theoretical models and most use bespoke approaches to measuring acceptability [20]. Therefore, it is evident that a more consistent approach to examining [20] and conceptualizing robot acceptability for OA and people with dementia needs to be developed. It may be helpful for this approach to include constructs that are pertinent to robot usage and includes factors used in gerontechnological models.

The empirical research will now be examined to identify the factors that impact robot acceptability and what is important to OA and people with dementia and is likely to impact their decision whether to use a robot.

*Collaborative and Humanoid Robots*

**Construct Definition**

Utilitarian Attitudes

Hedonic Attitudes

**Table 3.**

towards the robot.

the robot.

User Behaviour The user interacting with the robot.

*Constructs in the model of social robot acceptance [4].*

that has been described above.

model was tested by de Graaf [4] with respondents from an online questionnaire (n = 1248) of whom, 24.7% were aged over 60. This found that intention to use the robot increased when people believe: they have the skills to use the robot; that using it would be enjoyable; it would increase their status; when the robot would be less sociable; and cause less worry about privacy. However, only self-efficacy significantly impacted user intention to use the robot. De Graaf [4] concludes that the model is a useful guide to the identification of acceptance variables but that it needs further development and testing with other data sets. The Model of Social Robot Acceptance was used by Korblett [18] in the study examining telepresence robots

Personal Norms An individual's beliefs about using the robot, including privacy, trust, and attitude

Social Norms A user's evaluation of the social consequences of using the robot. This includes social

robot, including self-efficacy, previous experiences, prior expectations, personal

Includes perceptions of the robot's usefulness and how easy it is to use, including

Evaluation of emotions, including enjoyment and pleasure that might arise from using

influence, media influence, and impacts on status. Control Beliefs User's beliefs about resources, opportunities obstacles affecting the use of the

innovativeness, safety anxiety, and the cost.

Habit Behaviour resulting from the habitual use of the robot. User Intention The intention of the user to utilise and interact with the robot.

adaptability, embodiment, and robot personality.

One limitation of applying the Almere model and the Model of Social Robot Acceptance to OA and people living with dementia is that they do not take account of disability and the fact that OA and people with dementia may have additional needs concerning technology usage. It has been argued that everyday and psychosocial functioning are better predictors of robot acceptance than chronological age per se [16]. The usage of technology is impacted by a person's level of cognitive ability which affects their psychomotor speed, domain knowledge, and visual memory [11]. People with dementia may also experience declining touch sensitivity and less ability to execute accurate and discrete movements [6, 30]. Learning new technologies can also be difficult with symptoms of dementia such as reduced working memory, information processing ability and speed, and a lack of ability to disregard unwanted information [31]. Therefore, traditional models of technology acceptability and robot-specific models may be inadequate to explain the acceptance factors

Models of Gerontechnology acceptance have been developed which do consider

physical and cognitive health. Gerontechnology has been defined as electronic or digital products or services that can increase independent living and the social

The Senior Technology Acceptance Model (STAM) [30] was developed with data collected by personal interviews with community-dwelling people (n = 1012)

participation of OA [30]. Two examples of models are described below.

pertaining to some OA, particularly those with dementia.

*2.2.3 Acceptance models adapted for gerontechnology*

**22**

## **2.3 Factors that impact robot acceptance of people with dementia and older adults**

OA and people with dementia are motivated to use robots based on factors that impact them at the level of individual people and other factors that relate to people that are known to individuals (significant others), and wider society. This section of the chapter will firstly discuss how individual-level factors impact the motivation of potential and actual users to accept robots. It then addresses significant others and cultural factors.

The empirical studies included in this was obtained through searching eight databases using the following search terms: accept\*, dementia\*, Alzheimer\*, robot\*, "cognitive deficiency", elder\*, old\*, technology accept\*, user accept\*, attitude, social robots, assistive technology. Papers were selected if they were published between 2005 and 2021, they involved people with dementia, and/or adults over 65 years old, if they were in English, and focused on robot acceptability.

### *2.3.1 Individual factors that impact the motivation of OA and people with dementia to accept social robots*

Motivation to use robots is strongly related to their perceived usefulness. Social robots need to be relevant to the current unmet needs of potential users in order for robots to be perceived as useful [4, 14, 19, 28, 33–36]. Hebesberger et al. [36] aimed to investigate acceptance and the experience of using a humanoid robotic platform SCITOS, in a care institution in Australia. Data was collected in this 5-day pilot followed by a 15-day trial using semi-structured interviews and observations with people with dementia and their formal caregivers. They found that the robot was not perceived as useful and participants were ambivalent about the robot. Hebesberger et al. argues that acceptability is contingent on the robot meeting the specific needs of the end-users.

The identification of needs is complicated because OA and people with dementia may lack awareness of their unmet due to their cognitive difficulties [33], or because caregivers currently fulfil their needs [37] or because individuals are habituated to the challenges they are living with and don't perceive them as problems. But, people with dementia are very able to determine and articulate what they want and don't want and they know how they feel at any given moment [38, 39]. Potential users may also require information about the benefits of using the robot because they are new and unfamiliar technologies. Furthermore, the benefits of using the robot have to be real and clearly communicated to potential users [4, 28]. It is important to facilitate the development of realistic expectations to avoid users being disappointed if the expectations of using the robot are not met. This can result in a subsequent lack of acceptance [17].

In order to realise the benefit from the robot and to be motivated to use them, OA and people with dementia need to be capable of using a robot and to feel at ease doing so [16, 40]. This implies that there needs to be a user-technology fit [16] in terms of the robot capabilities to meet the users' needs. Thordardottir et al's., [17] systematic review that synthesised knowledge on the facilitators and barriers related to acceptance and the use of technology for people with cognitive impairment and caregivers found that PEOU was important and that acceptance was facilitated when technologies made low technical demands on users. A recent scoping review [41] also identified the robots need to be easy to use, to facilitate PEOU. This review included n = 53 studies and examined the use of barriers and facilitators affecting the implementation of zoomorphic robots by OA and people with dementia.

**25**

*Examining Social Robot Acceptability for Older Adults and People with Dementia*

Ha and Park [42] in their survey found that increasing age may present OA with more obstacles to technology acceptance, including cognitive, motor, and sensory deficits. They also found that PEOU negatively correlated with support, i.e. there was a greater need for support when PEOU was low. But, PEOU can be changed through the support of skilled facilitators. In a study that examined the effect of the robot MARIO on the resilience of people with dementia (n = 10) [6], it was found that all the participants needed facilitation and then they successfully used MARIO's touchscreen and interacted with the robot. Chan et al. [12] also found PEOU is important and that it can be improved through engagement with a robot. This randomised controlled trial investigated how participants' attitudes and beliefs towards a humanoid robot, Kabochan that resembles a 3-year-old boy, would change, and whether a change was affected by the amount and quality of engagement with the robot. People with dementia (n = 103) living in residential care facilities (n = 7) were allocated into a group to engage with Kabochan or to a control group who received usual care. An ABAB withdrawal experimental design was used with each phase lasting 8 weeks. The controlled group received usual care throughout the 32-week study. The experimental condition received usual care during the first baseline phase and the third phase (A phases). In the second phase, the robot-engagement group had 2 weeks of introductory sessions with an occupational therapist and they became familiar with Kabochan's interactive features and how it could be used and turned on and off. Then for six weeks and during the 8 weeks of the 4th phase participants could use Kabochan as they wanted independently 24 hours a day. This study found that PEOU changed significantly (p = −.042) and that it was related to the intensity of constructive behavioural engagement that

PEOU is related to self-efficacy which is an important outcome measure of positive psychology in dementia that is related to a feeling of being in control [43]. In order for users to feel at ease, they need to feel in control whilst they are interacting with robots as autonomous embodied devices. In the study described above Korblet [18] found that self-efficacy was mostly mentioned by the dementia group, and not by people with physical disabilities or visitors to the nursing home centre and that lack of self-efficacy led to participants not wanting to use the mobile presence robot again. Control requires a lack of fear, but also that the robot is reliable and can function properly to perform its tasks [10]. Conversely, technical robot problems and/ or a lack of robot capability can threaten acceptance [36, 44]. Frennert et al. [45] asked OA with moderate sensory and mobility impairments to state their preference for an ideal robot and OA (n = 7) who lived with mock-ups of these ideals for one week. They found that feelings of control were also linked to feelings of trust and the need for privacy. Trust may be particularly important for OA who may mistrust technology more than younger adults [40]. This was determined in an online survey conducted in Japan, with respondents (n = 100; aged 20–70). Trust is also related to the degree that a robot is autonomous and adaptive in its movement and behaviour. Korblet [18] found that trust did not play a role for participants in their research when the robot was not autonomous and it was only used for a few minutes on one occasion in a group setting. Rossi et al. [9] also found that trust and anxiety did impact technology acceptance. The more autonomous and adaptable a robot is, the more challenging it is to human comfort, the ability to trust the robot, and the need for perceived control of the robot. It may be that when robots are used over a longer period, their requirement for adaptivity may increase as they are required to respond in a larger number of human social settings. De Graaf [46] examined the acceptability of Karotz, a rabbit-like robot, in the homes of OA (n = 6) over three 10 day periods. They found that as time passed, participants wanted more control of the robot to maintain their privacy because the robot continued to remind

*DOI: http://dx.doi.org/10.5772/intechopen.98365*

participants had with the robot.

#### *Examining Social Robot Acceptability for Older Adults and People with Dementia DOI: http://dx.doi.org/10.5772/intechopen.98365*

Ha and Park [42] in their survey found that increasing age may present OA with more obstacles to technology acceptance, including cognitive, motor, and sensory deficits. They also found that PEOU negatively correlated with support, i.e. there was a greater need for support when PEOU was low. But, PEOU can be changed through the support of skilled facilitators. In a study that examined the effect of the robot MARIO on the resilience of people with dementia (n = 10) [6], it was found that all the participants needed facilitation and then they successfully used MARIO's touchscreen and interacted with the robot. Chan et al. [12] also found PEOU is important and that it can be improved through engagement with a robot. This randomised controlled trial investigated how participants' attitudes and beliefs towards a humanoid robot, Kabochan that resembles a 3-year-old boy, would change, and whether a change was affected by the amount and quality of engagement with the robot. People with dementia (n = 103) living in residential care facilities (n = 7) were allocated into a group to engage with Kabochan or to a control group who received usual care. An ABAB withdrawal experimental design was used with each phase lasting 8 weeks. The controlled group received usual care throughout the 32-week study. The experimental condition received usual care during the first baseline phase and the third phase (A phases). In the second phase, the robot-engagement group had 2 weeks of introductory sessions with an occupational therapist and they became familiar with Kabochan's interactive features and how it could be used and turned on and off. Then for six weeks and during the 8 weeks of the 4th phase participants could use Kabochan as they wanted independently 24 hours a day. This study found that PEOU changed significantly (p = −.042) and that it was related to the intensity of constructive behavioural engagement that participants had with the robot.

PEOU is related to self-efficacy which is an important outcome measure of positive psychology in dementia that is related to a feeling of being in control [43]. In order for users to feel at ease, they need to feel in control whilst they are interacting with robots as autonomous embodied devices. In the study described above Korblet [18] found that self-efficacy was mostly mentioned by the dementia group, and not by people with physical disabilities or visitors to the nursing home centre and that lack of self-efficacy led to participants not wanting to use the mobile presence robot again. Control requires a lack of fear, but also that the robot is reliable and can function properly to perform its tasks [10]. Conversely, technical robot problems and/ or a lack of robot capability can threaten acceptance [36, 44]. Frennert et al. [45] asked OA with moderate sensory and mobility impairments to state their preference for an ideal robot and OA (n = 7) who lived with mock-ups of these ideals for one week. They found that feelings of control were also linked to feelings of trust and the need for privacy. Trust may be particularly important for OA who may mistrust technology more than younger adults [40]. This was determined in an online survey conducted in Japan, with respondents (n = 100; aged 20–70). Trust is also related to the degree that a robot is autonomous and adaptive in its movement and behaviour. Korblet [18] found that trust did not play a role for participants in their research when the robot was not autonomous and it was only used for a few minutes on one occasion in a group setting. Rossi et al. [9] also found that trust and anxiety did impact technology acceptance. The more autonomous and adaptable a robot is, the more challenging it is to human comfort, the ability to trust the robot, and the need for perceived control of the robot. It may be that when robots are used over a longer period, their requirement for adaptivity may increase as they are required to respond in a larger number of human social settings. De Graaf [46] examined the acceptability of Karotz, a rabbit-like robot, in the homes of OA (n = 6) over three 10 day periods. They found that as time passed, participants wanted more control of the robot to maintain their privacy because the robot continued to remind

*Collaborative and Humanoid Robots*

**older adults**

cultural factors.

**2.3 Factors that impact robot acceptance of people with dementia and** 

OA and people with dementia are motivated to use robots based on factors that impact them at the level of individual people and other factors that relate to people that are known to individuals (significant others), and wider society. This section of the chapter will firstly discuss how individual-level factors impact the motivation of potential and actual users to accept robots. It then addresses significant others and

The empirical studies included in this was obtained through searching eight databases using the following search terms: accept\*, dementia\*, Alzheimer\*, robot\*, "cognitive deficiency", elder\*, old\*, technology accept\*, user accept\*, attitude, social robots, assistive technology. Papers were selected if they were published between 2005 and 2021, they involved people with dementia, and/or adults over 65 years old,

Motivation to use robots is strongly related to their perceived usefulness. Social robots need to be relevant to the current unmet needs of potential users in order for robots to be perceived as useful [4, 14, 19, 28, 33–36]. Hebesberger et al. [36] aimed to investigate acceptance and the experience of using a humanoid robotic platform SCITOS, in a care institution in Australia. Data was collected in this 5-day pilot followed by a 15-day trial using semi-structured interviews and observations with people with dementia and their formal caregivers. They found that the robot was not perceived as useful and participants were ambivalent about the robot. Hebesberger et al. argues that acceptability is contingent on the robot meeting the

The identification of needs is complicated because OA and people with dementia may lack awareness of their unmet due to their cognitive difficulties [33], or because caregivers currently fulfil their needs [37] or because individuals are habituated to the challenges they are living with and don't perceive them as problems. But, people with dementia are very able to determine and articulate what they want and don't want and they know how they feel at any given moment [38, 39]. Potential users may also require information about the benefits of using the robot because they are new and unfamiliar technologies. Furthermore, the benefits of using the robot have to be real and clearly communicated to potential users [4, 28]. It is important to facilitate the development of realistic expectations to avoid users being disappointed if the expectations of using the robot are not met. This can

In order to realise the benefit from the robot and to be motivated to use them, OA and people with dementia need to be capable of using a robot and to feel at ease doing so [16, 40]. This implies that there needs to be a user-technology fit [16] in terms of the robot capabilities to meet the users' needs. Thordardottir et al's., [17] systematic review that synthesised knowledge on the facilitators and barriers related to acceptance and the use of technology for people with cognitive impairment and caregivers found that PEOU was important and that acceptance was facilitated when technologies made low technical demands on users. A recent scoping review [41] also identified the robots need to be easy to use, to facilitate PEOU. This review included n = 53 studies and examined the use of barriers and facilitators affecting the implementation of zoomorphic robots by OA and people

if they were in English, and focused on robot acceptability.

*dementia to accept social robots*

specific needs of the end-users.

result in a subsequent lack of acceptance [17].

*2.3.1 Individual factors that impact the motivation of OA and people with* 

**24**

with dementia.

participants about their schedule and health-promoting activities in the presence of guests. Furthermore, how acceptable individual users find a robot depends on the individual person, the purpose of the robot, and the context [14].

The personality of the user, their interests, and values impact their response to a social robot and their perception about its social presence [9]. Rossi et al. found that having the personality trait of 'openness to experience' positively impacted the human-robot interaction. This study used a prototype of the humanoid robot Pepper to investigate if personality traits and user's empathy (a feature of personality) impacted the acceptance of a robot-led cognitive test. Participants were OA (n = 21; aged = 53–82 with average of 61). Acceptance of the robot was assess using, the UTAUT constructs and a psychologist evaluated personality traits, empathy, using the NEO Personality Inventory-3 (NEO-PI-3) [47]; the Empathy Quotient (EQ ) [48]. After the psychological evaluation, the robot administered to participants psychometric tasks of the MoCA [49], and in a second task the person performed activities whilst Pepper monitored them. Dialogues of the participantrobot conversation and videos of the interaction were captured by Pepper. Rossi et al. found that empathy correlates with the amount the user perceived that the robot was sociable.

People with dementia can experience volatility in their mood and difficulty regulating their emotions [39]. Variations in mood are likely to impact robot acceptability [28]. When MARIO was being used by people with dementia, who had moderate and severe dementia, facilitators had to support participants to enable them to be ready to use the robot [6]. If participants were disorientated in time or space, facilitators had to acknowledge their perception of reality at that time and help the participants to deal with whatever was causing their anxiety. For example, at the start of one session with the robot, the participant welcomed MARIO and the facilitator into her room in the nursing home, but she was concerned, believing erroneously that a person had been into her house without her permission. This participant was not ready to use MARIO until she was calmed as a result of talking with the facilitator and being reassured that all was well [6].

OA and people with dementia are also motivated to use a robot if they find that interacting with it is enjoyable [19, 41, 50]. Novelty effects may enhance the enjoyment of robot usage initially, but these can decrease over time [46, 51]. How robots can be used to sustain a person's enjoyment and its relevance to their needs and capabilities, are discussed below. But first, the factors involving significant others and wider society that impact an individual's motivation to accept robots will be introduced.

#### *2.3.2 Factors that involve significant others and wider society*

The perceptions of significant others, particularly caregivers can have a substantial impact on robot acceptability. Significant others, as social influences, are strong predictors of the adoption of home healthcare robots [52] and the ITU the Kompai robot [53]. Significant others can enable the usage of robots through encouragement and facilitation [6, 53] or they can impede a robot's implementation into care settings [36, 44]. There is substantial evidence that professional health and social caregivers may have negative preconceptions about the use of robots for OA and people with dementia [44]. Caregivers may be concerned about the compatibility of robots with their work processes [41]. Casey et al. [8] explored the perceptions and experiences of using MARIO with people with dementia (n = 38), relatives/carers (n = 28), formal carers (n = 28) and managers (n = 13) in UK, Italy, and Ireland. They found that although MARIO was positively received by most of the participants, some formal care workers voiced concern that robots might replace care staff.

**27**

discussed.

**with dementia**

*Examining Social Robot Acceptability for Older Adults and People with Dementia*

gies is impacted by OA subjective interpretations of the technologies.

**2.4 What is likely to increase robot acceptability for OA and people** 

Robot acceptance is increased through personalisation of the robot, to make it meet the needs and preferences of the user and support their personhood [7, 14, 17, 44]. Personhood has been defined as 'a standing or a status that is bestowed on one human being by another, in the context of relationship and social being' [59] p.8. For example, the acceptability of the robot Maltida was increased due to personalisation that supported personhood, through the robot using human-like emotive expressions and accounting for the user's disabilities [7]. Other ways to support personhood and increase the acceptability of a robot is if the robot enables the OA and person with dementia to feel empowered, respected, and able to participate in activities that are meaningful to them [6, 10]. It is well recognised in current discourses about dementia care, that people with dementia do not experience a 'loss of self' as dementia progresses [39]. Whilst a person's identity, and personality may change, a person's central being, core values remain [39] and people with dementia maintain the potential to adapt and grow [38]. It is not enough to discern the requirements of the robot user on one occasion and personalise the robot. To ensure ongoing acceptance of the robot, there must be ongoing review and adaptation of the robot as the person uses it over time. The investigation with MARIO described above found that the creation of meaningful activity was made possible through a skilled facilitator and person with dementia working in partnership to use the robot in a way that was meaningful to the user [6]. Facilitators learned about the user's preferences and used the robot creatively in ways that were not identified before usage, even though MARIO was initially

Another factor that impacts the level of robot acceptability for individuals, significant others, and wider society, is the familiarity of robots as a technology [44]. It is argued that people need time to learn about and become comfortable with robots as a technology [17, 44, 51] and that this may happen when robots are more prevalent. The level of robot acceptability will vary between cultures and over time, but as robots become more available and diffuse in society it is not certain that acceptability will increase [28]. Bishop et al. found that higher robot familiarity had a negative impact on attitudes and behaviours and that familiarity was linked to heightened awareness of a robot's limitations. How robots need to be design, developed, and deployed used in order to facilitate their acceptability will now be

Human behaviour is impacted by a person's perceptions about what other people think, including thoughts related to negative ageing stereotypes that are prevalent in society [54]. People are motivated to act because they want to project a self-image that they are healthy and independent, both to themselves and other people [33, 55, 56]. The acceptance of robots by people with dementia and OA is impacted as potential users may fear being stigmatised and labelled as physically or psychologically vulnerable, dependent, in decline, or lonely if they use a robot [34, 44, 53, 56]. Dudek et al. introduced OA (n = 28) in good subjective physical and cognitive health to PLEO, a robot shaped like a dinosaur, and then assessed robot acceptance using an Emotional Attachment Scale (EA-Scale) developed by Thomson et al. [57] and the Comfort from Companion Animals Scale developed by Zasloff [58]. Then they accessed the participants' actual and ideal self and subjective robot user image. They found that participants stigmatised OA robot users and that OA may construct a more negative user image than their own self-image, in order to maintain their own positive self-image. Dudek et al. concluded that acceptance of new technolo-

*DOI: http://dx.doi.org/10.5772/intechopen.98365*

#### *Examining Social Robot Acceptability for Older Adults and People with Dementia DOI: http://dx.doi.org/10.5772/intechopen.98365*

Human behaviour is impacted by a person's perceptions about what other people think, including thoughts related to negative ageing stereotypes that are prevalent in society [54]. People are motivated to act because they want to project a self-image that they are healthy and independent, both to themselves and other people [33, 55, 56]. The acceptance of robots by people with dementia and OA is impacted as potential users may fear being stigmatised and labelled as physically or psychologically vulnerable, dependent, in decline, or lonely if they use a robot [34, 44, 53, 56]. Dudek et al. introduced OA (n = 28) in good subjective physical and cognitive health to PLEO, a robot shaped like a dinosaur, and then assessed robot acceptance using an Emotional Attachment Scale (EA-Scale) developed by Thomson et al. [57] and the Comfort from Companion Animals Scale developed by Zasloff [58]. Then they accessed the participants' actual and ideal self and subjective robot user image. They found that participants stigmatised OA robot users and that OA may construct a more negative user image than their own self-image, in order to maintain their own positive self-image. Dudek et al. concluded that acceptance of new technologies is impacted by OA subjective interpretations of the technologies.

Another factor that impacts the level of robot acceptability for individuals, significant others, and wider society, is the familiarity of robots as a technology [44]. It is argued that people need time to learn about and become comfortable with robots as a technology [17, 44, 51] and that this may happen when robots are more prevalent. The level of robot acceptability will vary between cultures and over time, but as robots become more available and diffuse in society it is not certain that acceptability will increase [28]. Bishop et al. found that higher robot familiarity had a negative impact on attitudes and behaviours and that familiarity was linked to heightened awareness of a robot's limitations. How robots need to be design, developed, and deployed used in order to facilitate their acceptability will now be discussed.
