Impact of Anthropomorphic Design on User Sentiment and Sustained Use Intention towards Household Healthcare

Abstract

Although the household healthcare system is one of the cutting-edge application areas of anthropomorphic design, it remains to be further explored whether anthropomorphism is always effective. This article focuses on the context of aging-friendly household healthcare services and explores how anthropomorphic design affects users’ sustained use intention through sentiment feedback mechanisms. With the help of questionnaire surveys, 511 valid samples were randomly collected for empirical analysis and hypothesis testing. The results showed that positive interactions, cultural backgrounds, and appearance could enhance users’ perception of anthropomorphism from large to small. In addition, the positive (negative) sentiment of users plays a positive (negative) full mediating role in the relationship between anthropomorphic design and sustained use intention, and user technology anxiety moderates such relationships. That is, compared to low-level technology anxiety, in high-level states, anthropomorphic design for household healthcare systems and products has a weaker (stronger) positive (negative) effect on sustained use intention through positive (negative) emotions.

1. Introduction

With the rapid development of artificial intelligence (AI), intelligent products with anthropomorphic features are increasingly widely used in wise information technology in medical (WITMED) and household healthcare systems [1]. WITMED, as a new medical model that improves quality and efficiency, brings obvious benefits to both users and the medical industry, and many household healthcare systems have emerged as a result [2]. However, there are still some issues that need to be addressed in the current user experience process, such as users’ mixed sentiments and attitudes towards anthropomorphic design [3]. How to resolve users’ differentiated perceptions and enhance their sustained use intention is a crucial issue.

The anthropomorphic features of products are mainly divided into two categories: functional and emotional [4]. Functional smart products emphasize improving efficiency, reducing costs, and striving to improve the quality of physical therapy. Emotional smart products focus on enriching users’ psychological experiences and improving their mental health through natural dialogue and accurate understanding of emotional needs. The differences in the anthropomorphic features of intelligent systems can divide users’ attitudes toward their sustained use intention [5]. Existing research has shown that in retail service settings, the anthropomorphic features of intelligent robots affect consumer behavior, which in turn affects their willingness to use them [6]. In addition, scholars have also pointed out that the usefulness, social competence, appearance, and other anthropomorphic features of intelligent robots could affect customer acceptance [3,7,8,9]. However, there are currently few studies on the service effects of anthropomorphic design in household healthcare systems and products, especially the impact of anthropomorphic features on users’ feedback of sentiment.

Sentiment towards the anthropomorphic design of the system could be divided into two dimensions: cognitive and emotional [10]. In terms of the cognitive dimension, users are more concerned about system transparency [11]. AI technology represented by deep learning is very complex and inherently opaque, and it is often seriously affected by failures caused by even small disturbances in the environment, showing the uncertainty and vulnerability of AI compared to other technologies, which directly affects users’ anxiety and doubts about smart systems [12]. In terms of emotions, anthropomorphism is a core factor in the acceptance of smart systems, especially in the AI application field represented by service robots [3]. Anthropomorphic service robots are considered by humans to have certain social skills and can produce experiences of empathy. However, over-anthropomorphic robots may also produce the uncanny valley effect, causing discomfort or even fear in users [13].

In summary, the issue of sentiment is a core factor in human acceptance of intelligent healthcare systems. In empirical research, the sentiment feedback mechanism in an AI context is divided into positive and negative dimensions [14]. However, due to the diverse application areas of AI, such as autonomous driving, and robotics, there is a strong heterogeneity in the equipment used, application scenarios, and user identity characteristics of AI technology [3,9,10]. This leads to potentially divergent results in empirical research and prevents general conclusions about user sentiment in intelligent systems from being drawn. Therefore, this study will focus on the household healthcare application scenario serving the elderly population, investigating how anthropomorphic design affects user sentiment. This may greatly help predict the acceptance and sustained use intention of different user groups for intelligent healthcare systems. The theoretical value of this article is to propose a theoretical model for the sustained use intention of household healthcare systems and products, revealing their sentiment mechanism and boundary effects.

2. Theoretical Background and Hypothesis Development

2.1. Connotation of Anthropomorphic Design

From a psychological perspective, anthropomorphism refers to the process of assigning human characteristics such as traits, intentions, and psychological states to non-human individuals, making them appear as living, thinking beings [15]. This is a process of inductive reasoning for non-human subjects, and the basic cognitive operations involved in this reasoning are the same for anthropomorphism as for any other reasoning [7]. It includes acquiring knowledge, activating or evoking stored knowledge, and applying the activated knowledge to a given goal.

According to the anthropomorphic three-factor theory proposed by Epley et al. [16], people’s perception of human-like cues exhibited by non-human objects is one of the important antecedents of anthropomorphism. Elicited agent knowledge is the foundation for anthropomorphism, and effecting motivation and sociality motivation are the driving forces behind it. The three factors complement each other and are indispensable, jointly constituting a complete anthropomorphic process.

With the development and progress of the times, designers have gradually shifted their focus from prioritizing functionality to the psychological feelings of users. Donald A. Norman proposed a three-level theory of emotional design from the perspective of cognitive psychology, namely the instinctual level, the behavioral level, and the reflective level. By designing anthropomorphic emotional experiences for the instinctual, behavioral, and reflective levels, designers could establish a matching relationship with users during the process of anthropomorphism decoding, meet users’ multi-level needs, and optimize the process of anthropomorphism [17].

It could be seen that the anthropomorphic design of the household healthcare service system is also reflected in its appearance, interactive feedback methods, and cultural background. In addition to the kinematic and structural advantages brought by humanoid robots, anthropomorphism could also stimulate special psychological responses and affect the user’s interaction process from an emotional level. The essence of anthropomorphic design is emotional design [18], so whether it is applied in appearance, interaction, or cultural design, it is all about enhancing its emotional interaction with people to provide a better user experience. Therefore, the following hypothesis is proposed.

Hypothesis 1.

Appearance, interaction, and cultural background are the three second-order variables of anthropomorphic design for household healthcare systems and products. The positive performance of these three elements will enhance the user’s perception of anthropomorphism.

2.2. Anthropomorphism and User Sentiment

Mainstream research has shown that anthropomorphism increases consumers’ positive experience and alleviates negative emotions. For example, deformed foods that appear to be smiling could trigger positive emotions in consumers [14]. Anthropomorphic communication alleviates consumers’ vigilance when they come into contact with brands [19]. In the process of online marketing, companies using anthropomorphic communication can compensate for the lack of presence in online consumption, reducing consumers’ perception that brands are impersonal and inhuman, and enhancing customer happiness [20]. When consumers receive personalized marketing using anthropomorphic communication with low information sensitivity, they are more likely to regard personalized marketing as sharing between friends rather than having cold sales purposes, resulting in more positive feelings and reducing privacy concerns [21]. The use of anthropomorphism in advertising services could also weaken consumers’ perception of the hidden motives of advertisers and reduce their suspicion of advertising information [5]. What is more, research has found that anthropomorphic products are perceived as human-like subjects, which bear some responsibility for consumers’ indulgent consumption, reducing consumers’ perception of control and responsibility for their own behavior [22], and thereby reducing the psychological barriers to purchasing “hedonic products”.

Anthropomorphism often triggers positive emotions, such as excitement, curiosity, and liking, which are the result of interacting with robots. Numerous studies have shown that different populations, including children and the elderly, have a high level of interest and acceptance towards robots [18,23]. For example, by examining the anthropomorphic features of participants towards different service robots, it was found that more anthropomorphic features of robots are associated with more emotional trust and a more enjoyable interaction experience [9,23]. In addition to the anthropomorphism of robots, in the field of automatic driving, Waytz et al. [24] investigated the effect of anthropomorphism by giving autonomous vehicle sounds and names. The results indicate that compared to non-anthropomorphic cars, anthropomorphic cars are more trusted, thus proving the positive power of anthropomorphism.

However, a few studies have verified that there are still certain risks in the application of anthropomorphic design. For example, anthropomorphic websites seem to have “motivation” and “intent”, which have triggered consumers’ concerns about being followed, tracked, and even manipulated by real or imagined others [25]. Consumers with low social needs also show higher privacy concerns about anthropomorphic websites [26].

In summary, mainstream research suggests that anthropomorphic design could increase users’ positive emotions and reduce negative emotions. A small number of studies have verified that anthropomorphism triggers negative emotions. Therefore, further empirical study is necessary to explore the reasons for inconsistent conclusions. Hence, we propose the following hypothesis.

Hypothesis 2.

Anthropomorphic design for household healthcare systems and products will enhance (weaken) users’ positive (negative) sentiment.

2.3. User Sentiment and Sustained Use Intention

In research on the adoption of new technologies, user sentiment and emotional trust are important factors used to explain the acceptance of technology. Previous studies have found that trust has a positive impact on perceived usefulness and enjoyment, thereby increasing users’ sustained use intention towards the new technology of virtual agents. In research on the relationship between trust and perceived ease of use, Mostafa and Kasamani [27] found that perceived ease of use, compatibility, and social influence significantly increased users’ initial trust in chatbots with AI features. Whether it is autonomous driving [28], robots [9], or other information systems with AI features [29], a large number of studies have shown that increased trust or positive sentiment could significantly enhance users’ sustained use using intention.

However, a small number of studies suggest that the negative emotions triggered by anthropomorphic design may lead to negative perceptions. For example, according to the uncanny valley theory, the mystery of anthropomorphic agents will reduce consumer attitudes [13]. Moreover, an anthropomorphic brand that makes mistakes will trigger more negative evaluations and affect consumers’ sustained use intention [19]. Overall, based on the analysis of previous empirical literature, this study suggests that:

Hypothesis 3.

Positive (negative) user sentiment will enhance (reduce) the sustained use intention towards anthropomorphic household healthcare systems and products.

Hypothesis 4.

User sentiment plays a mediating role in the relationship between anthropomorphic design and sustained use intention.

2.4. Moderating Effect of User Technology Anxiety

High-tech anxiety may reduce customers’ willingness and possibility to use smart systems. When customers are not proficient enough in the operation of smart systems, they may experience high levels of technical anxiety and nervousness, thus avoiding interaction with smart systems. This in turn weakens their willingness to continue using smart systems [12]. Conversely, when customers have low levels of technical anxiety, they will feel relaxed and confident in using smart systems, resulting in enthusiasm and enhancing their willingness [30].

Furthermore, emotional consistency theory research shows that people will judge the quality, positive or negative, and risk level of a situation based on their emotional state, then make corresponding decisions and behaviors based on these judgments [31]. Users’ goal analysis is affected by incidental feelings, and poor experience perception will lead to consumer evaluation bias. According to the above discussion, when users are in high-tech anxiety, it is easy to trigger negative sentiment. Therefore, when users’ technical anxiety increases, the positive effect of anthropomorphic design of household healthcare systems with positive (negative) sentiment as a mediating variable on sustained use intention may weaken (strengthen).

Hypothesis 5.

Compared to low-level technology anxiety, in high-level states, anthropomorphic design for household healthcare systems and products has a weaker (stronger) positive (negative) effect on sustained use intention through positive (negative) emotions.

3. Methods

3.1. Participants and Procedure

This article aims to explore how anthropomorphic design affects users’ sustained use intention through sentiment feedback mechanisms. In order to ensure the pertinence and depth of the research, a household healthcare application scenario serving the elderly population was selected, with elderly individuals and their adult children as the empirical research subjects. Moreover, in order to make it easier for the participants to understand the research purpose, necessary descriptive materials about the service model, product examples, and commonly used functions of the household healthcare system were placed on the front page of the survey text.

The scope of the household healthcare system mentioned in our material focuses on remote care for chronic diseases and elderly patients, with three main functions. The first is an intelligent medication reminder that automatically prompts medication time, contraindications, and remaining medication amount. The second is a comfort system that includes chat and control of home audio and video systems. The third is real-time monitoring of physiological data such as heart rate and blood pressure in elderly patients, and sending out alerts when necessary. In addition, the interpretation of anthropomorphism is mainly reflected in its appearance, interaction methods, and language style.

We integrated a questionnaire and conducted a simple random sampling survey to collect samples through a crowdsourcing platform in China (www.wjx.cn, accessed on 15 January 2024) from January to February 2024. Using a simple random sampling method could improve the generality of the sample and enhance the authenticity and quality of the questionnaire survey [32]. The questionnaire developed by the authors is divided into three parts. The first part is the introduction and reading materials, where the participant could know the survey procedure and disclosure of a data confidentiality statement. When the participant fully understands and agrees to participate in the survey, they can enter the subsequent steps. The second part is the scale measurement questions, and the third part is demographic information. Before formal investigation, five graduate students independently proofread the text materials to ensure the readability of the questionnaire.

Participants received a reminder of data confidentiality again, verbal gratitude, and monetary rewards from the researchers after submitting the questionnaire. Finally, this study collected 531 samples, excluding 12 samples that only filled in the first part (i.e., refused to participate) and 8 samples that had incorrect options in the verification questions (1 + 2 + 3 = 6), and finally retained 511 valid samples. The key demographic information (

Table 1. Several key demographic information.

Demographic	Items	N	Percentage
Gender	Male	238	46.58
	Female	273	53.42
Age	18–35	26	5.09
	36–40	63	12.33
	41–50	97	18.98
	51–55	126	24.66
	56–60	150	29.35
	60–73	49	9.59
Objects	For themself	210	41.10
	For parent or grandparent	301	58.90

) shows that the sample distribution is relatively uniform and could reflect the overall characteristics of the target consumers.

3.2. Variable Measurement

The measurement of the core variables was adapted from previous studies using multiple-item scales. To assess various independent concepts, a five-point Likert scale was used for scoring, with respondents scoring from 1 (strongly disagree) to 5 (strongly agree) to express their level of agreement with the statements in the items.

By adapting the scales of Xie et al. [26], Eyssel et al. [33], and Puzakova et al. [19], we measured the appearance, interaction, and cultural background variables in anthropomorphic design, including a total of 9 items. The user sentiment variable including 6 items was adapted from the scales by Zhang et al. [18] and Puzakova et al. [19]. Meanwhile, the measurement of the sustained use intention variable refers to the scale from Gupta et al. [34] and Lee and Park [8] for evaluation, including a total of three items. As for user technology anxiety, we adapted a scale from Meuter et al. [35], which focuses on users’ behaviors and attitudes during technology use, including six items. In addition, to more comprehensively explore the impact of various factors on our conceptual model, this article also considered gender, age, and education as important control variables to ensure the depth and breadth of the research.

Table 2. Measurement references for Likert scale.

Variable	Measurement by Likert Scale
Anthropomorphic design
-Appearance	Has a rounded and full appearance
	It looks cute
	It is easy to associate with the appearance of ordinary people
-Interaction	Actions and behaviors similar to those of humans
	Sounds and intonations similar to those of humans
	Language is natural, as if chatting with a real person
-Cultural background	The language style is consistent with the living environment
	The appearance image is consistent with the living environment
	The name is consistent with the living environment
User sentiment
-Positive	Feel cared for
	Inner warmth
	Pleasure
-Negative	Embarrassing
	Doubt its effectiveness
	Noisy feeling
Sustained use intention	Will use it frequently
	Will continue to use it instead of stopping using it
	Will recommend others to use it
Technology anxiety	Difficult to understand its technical problems
	Concerned about the use of such technology
	Worried that they might damage it in some way
	Confused by a bunch of terms
	Avoided using it because I am not familiar with it
	Dare not use it for fear of making irreversible mistakes

summarizes the measurement references for the above variables.

4. Results

4.1. Reliability and Validity

To ensure the reliability and validity of the study, this article conducted a reliability and validity test on the scale based on random sampling data in SPSS.

Table 3. Reliability and validity of each variable.

Variable	Loadings	Cronbach’s α	CR	AVE
Anthropomorphic design		0.906	0.9209	0.5649
Appearance	0.808; 0.743; 0.786	0.876	0.8227	0.6076
Interaction	0.702; 0.802; 0.779	0.864	0.8057	0.5809
Cultural background	0.686; 0.697; 0.750	0.792	0.7544	0.5063
Positive sentiment	0.658; 0.836; 0.852	0.687	0.8281	0.6193
Negative sentiment	0.776; 0.704; 0.861	0.696	0.8252	0.6130
Sustained use intention	0.781; 0.825; 0.785	0.765	0.8395	0.6356
Technology anxiety	0.840; 0.841; 0.774; 0.763; 0.750; 0.715	0.934	0.9039	0.6113

shows the results of the reliability and validity. The results show that Cronbach’s α are all greater than or close to 0.7, and the combined reliability is greater than 0.7, indicating that the items have high consistency. In addition, the AVE values of each variable are all greater than 0.5, indicating that the convergent and discriminant validity of the scale meet the requirements.

To evaluate the reliability and validity of the multi-item structure in the model, confirmatory factor analysis was conducted. The fit index was χ² = 281.103, df = 160, RMSEA = 0.054, CFI = 0.939, TLI = 0.928, indicating good fit. This article measured discriminant validity by comparing the square root of the AVE of individual structures with the correlation between all variables. As shown in

Table 4. Correlation between variables.

Variable	Mean	SD	1	2	3	4	5
1: Anthropomorphic design	3.982	0.685	0.752
2: Positive sentiment	3.688	0.711	0.617 **	0.787
3: Negative sentiment	3.433	0.684	0.558 **	0.645 **	0.783
4: Sustained use intention	3.113	0.786	0.293 **	0.494 **	0.689 **	0.797
5: Technology anxiety	3.459	0.889	0.224 **	0.523 **	0.516 **	0.612 **	0.782

Note: ** p < 0.01.

, the square root of the AVE of each construct was greater than the highest correlation between all variables, indicating satisfactory discriminant validity.

Multicollinearity among variables may affect the regression results of the model. In this article, it was found through the Harman single-factor test that: (1) there were five factors with eigenvalues greater than 1, exceeding 1; (2) the variance contribution rate of the first factor was 27.914%, which was lower than 30%. Therefore, there was no serious common method bias in this article.

4.2. Hypothesis Testing

Using AMOS 23.0 and SPSS 22.0 to conduct hypothesis testing, and the results are summarized in Figure 1 and

Table 5. Regression analysis results of hypothesis testing.

Path	Model 1 (DV: PS)		Model 2 (DV: NS)		Model 3 (DV: SUI)
	β	CI	β	CI	β	CI
Direct effect
AD	0.642 ***	0.550, 0.733	0.540 ***	0.444, 0.636	0.109	−0.235, 0.017
PS					0.081 *	0.280, 0.442
NS					−0.454 **	−0.914, −0.006
Mediating effects
AD⟶PS					0.152 ***	0.096, 0.223
AD⟶NS					−0.022 *	−0.098, −0.035
Moderating effects
PS × TA					−0.034 *	−0.144, −0.076
NS × TA					0.281 **	0.166, 0.397

Note: * p < 0.05; ** p < 0.01; *** p < 0.001; CI represents the confidence interval, which includes the lower limit of confidence interval (LLCI) and upper limit of confidence interval (ULCI).

. The results of the direct effect test by AMOS show that the positive performance of the three second-order variables in anthropomorphic design, interaction (β = 0.891, p < 0.001), cultural background (β = 0.840, p < 0.001), and appearance (β = 0.808, p < 0.001 could enhance users’ perception of anthropomorphism from large to small, supporting H1. Among the direct effects, the anthropomorphic design has a significant promoting effect on enhancing positive sentiment (β = 0.642, p < 0.001) and weakening negative sentiment (β = 0.540, p < 0.05), supporting H2. In addition, the positive sentiment has a significant positive effect (β = 0.081, p < 0.001) on sustained use intention, while negative sentiment has a significant negative effect (β = −0.454, p < 0.01), supporting H3. However, in the main effect model, the direct effect of anthropomorphic design on sustained use intention is not significant (β = 0.109, p = 0.436 > 0.05).

To further explore the mediating and moderating effects, we conducted a moderated mediation test by model 14 in SPSS PROCESS. The results (Table 5) showed that both positive (β = 0.152, p < 0.001) and negative (β = −0.022, p < 0.05) sentiment played a fully mediating role in the path relationship between anthropomorphic design and sustained use intention, supporting H4. As for technology anxiety, it played a significant moderating effect in the above model mechanism. That is, it weakens the positive impact of positive sentiment on sustained use intention (β = −0.034, p < 0.05) and enhances the negative impact of negative sentiment on sustained use intention (β = 0.281, p < 0.01), establishing H5.

5. Conclusions and Discussion

5.1. Conclusions

This article explored how anthropomorphic design affects users’ sustained use intention through sentiment in aging-friendly household healthcare services through empirical research methods. The results showed that interaction, cultural background, and appearance, as the three second-order variables of anthropomorphic design, could enhance users’ perception of anthropomorphism by positive performance from large to small. That is, anthropomorphic interactive design will stimulate stronger perceptions and preferences of anthropomorphism compared to cultural background and appearance design.

In addition, positive (negative) emotional trust plays a positive (negative) mediating role in the relationship between anthropomorphic design and sustained use intention, and user technology anxiety moderates the above main effects. That is, compared to low-level technology anxiety, in high-level states, anthropomorphic design for household healthcare systems and products has a weaker (stronger) positive (negative) effect on sustained use intention through positive (negative) emotions.

5.2. Theoretical Implication

First, this study integrates inconsistent conclusions about the impact of anthropomorphism on consumer emotions, and identifies situations where the negative effects of anthropomorphism are likely to occur (household healthcare system), deepening the application of the uncanny valley theory in the field of marketing and industrial design. On the one hand, anthropomorphism would indeed improve the consumer experience and stimulate positive emotions. On the other hand, the uncanny valley theory argues that anthropomorphism will trigger negative emotions in specific situations. Using empirical analysis to further refine the findings, it is found that in the household healthcare service context, while anthropomorphism brings positive emotions to users, it does not reduce but rather increases negative emotions such as embarrassment and concern. The conclusion of this study acknowledges the positive effects of anthropomorphism and clarifies the neglected issue of the effectiveness of anthropomorphism in reducing negative emotions in the household healthcare service context.

Second, we have also clarified the moderating effect of technology anxiety. Research has shown that although anthropomorphism could effectively enhance users’ positive emotions towards healthcare systems, this is only possible under low levels of user technology anxiety. Human user experience is divided into pragmatic and hedonic experiences. Similarly, household healthcare systems also differ in their functional capabilities between pragmatic and hedonic functions. For household healthcare systems related to health, the demand for pragmatic functions is greater during emergency periods (such as dying). However, in daily wellness services, due to constant contact with users, in addition to fulfilling pragmatic functions, it is also necessary to satisfy the hedonic functions of physical and mental care to achieve pleasant communication with users. At this point, it is necessary to consider the issue of system anthropomorphism. Empirical results show that for users with high levels of technology anxiety, anthropomorphic design brings them lower positive effects and more obvious negative consequences, confirming previous research conclusions.

5.3. Management Implication

Firstly, companies should make good use of anthropomorphic methods based on the characteristics of the object and the psychological characteristics of the audience, considering factors such as product attributes and cultural customs, and designing appropriate anthropomorphism that conforms to public values and legal ethical constraints. That is a relatively safe and effective strategy when external risks cannot be predicted. In addition, companies should also consider the psychological needs of target consumers and convey emotions through anthropomorphic elements. The anthropomorphism used in the home should be designed to be as cute and warm as possible, reducing the degree of anthropomorphism in appearance, avoiding the risk of the uncanny valley, and making the family atmosphere relaxed and pleasant. For example, in daily life, more “warm” anthropomorphic products are introduced to elderly people who have been living alone for a long time, highlighting brand companionship and support.

Secondly, companies should adopt different anthropomorphic design strategies for different use scenarios to avoid negative emotions that may be brought to users. Especially with the development of AI technology, robots have penetrated into consumers’ daily life scenarios, not only weakening users’ embarrassment and anxiety, but also dealing with privacy concerns caused by high-capacity perception robots, reducing or overcoming risks or hidden dangers caused by the trend of precision marketing.

Finally, although relying on technology can increase consumer interaction with brands, whether customer traffic may be efficiently converted into product sales and corporate benefits requires companies to scientifically design anthropomorphic marketing plans. For example, considering the design of anthropomorphic styles in the family cultural atmosphere, in the face of users with such collectivist values, the service system should adopt an anthropomorphic design that is popular among most members.

5.4. Social Implication

The global aging issue has attracted widespread attention from society. Considering the physiological and psychological changes that have occurred among the elderly, especially those who live alone, corresponding household healthcare systems should be designed for the elderly population, and sustainable design concepts should be integrated into them. This would aim to improve their quality of life and happiness, and enable them to enjoy the convenience changes brought by the intelligent era. Starting from the concept of sustainable design, designing a household healthcare system suitable for the elderly is not only to meet the needs of elderly care, but also to promote the sustainable development of the elderly industry.

On the other hand, a household healthcare system that is suitable for the elderly should not only focus on the emotional interaction between the elderly and the service system, but also pay attention to social value. Therefore, the specific design work should follow the principles of practicality, usability, humanization, and anthropomorphism, comprehensively considering the characteristics of the elderly and social value orientation. For example, ensuring that the interaction process has clear and distinct information feedback language and actions, coordinated and easily recognizable color combinations, and clear operational logic, so that elderly people can easily recognize, remember, and operate products, which not only satisfies their emotions, but also promotes their physical and mental health. In addition, with the help of anthropomorphic design, it may also be possible to bring the distance between parents and younger generations closer, allowing the care of younger generations to be integrated into every aspect of the elderly’s life.

5.5. Limitations and Future Research

There are several limitations in this article that need further research in the future. Firstly, this study used user sentiment as mediating variables and anthropomorphic design as antecedent variables for empirical analysis. However, the self-regulation mechanism of sentiment or emotions also includes other mediating variables such as cognitive trust, reliability, and embodied cognition. Therefore, it may be possible to specifically discuss cognitive trust or emotional trust in the future. Secondly, regarding research methods, current research on AI interaction and collaboration is still dominated by empirical research using similar methods to this article, with few neuroscience experiments being conducted. In the future, longitudinal experiments could be conducted to collect relevant neuroscience experimental data and obtain more comprehensive results.