The Robotic Multi-Care Network: A Field Study of a “Robot Grandchild” in South Korea

Abstract

Technical solutions are being presented for the healthcare and welfare of the aging population. One example is Hyodol, a robot developed for older adults living alone in South Korea. By offering various services such as religious chanting, dementia-prevention quizzes, and daily medication reminders, Hyodol is expected to serve as a companion for lonely older adults. This paper analyzes how the robotic care program for older adults is operating within the Korean public welfare system. Based on our fieldwork at regional welfare institutions and older adults’ homes, we show that the robots along with its monitoring system, older adults, institutional managers, caregivers, company staff, and family are forming what we call a “robotic multi-care network.” Within this network, the elderly users cultivate their own ways of building relationships with the robot, some perceiving it as a “grandchild” while others view it as a medium to connect with caregivers. The introduction of the care robot at the welfare institutions does not make their elderly care work unmanned, nor does the robot substitute for human caregivers. Instead, it displaces and redistributes the caregivers’ tasks and responsibilities, leading to multiple eldercare practices—tactile, digital, proximate, remote.

Keywords:

• Older adults
• care robot
• robotic multi-care network
• South Korea
• Hyodol

1 Introduction

Declining birth rates and aging populations are a global problem, but it is more serious in South Korea than in any other country. In 2020, South Korea recorded the lowest fertility rate among the Organisation for Economic Co-operation and Development (OECD) countries for seven consecutive years. There were only 0.84 births per woman, a rate much lower than those of its neighbors Japan (1.33) and China (1.70) (OECD 2022).¹ In this rapidly aging population, an increasing number of older adults are living alone. In 2020, 35.1% of the elderly population was reported as one-person households (Yonhap 2021). South Korea’s Ministry of Health and Welfare (MOHW) estimates that the number of older adults living alone will more than double by 2035 (Ministry of Health and Welfare 2018: 1). These demographic trends raise urgent questions about elderly care: How should older adults living alone be taken care of? Are there enough people for this work of caring?

Against this backdrop, robots and artificial intelligence technologies have emerged as useful measures to respond to the problem of elderly care in an aging society. In recent years, the “AgeTech” industry has expanded considerably, backed by the Korean government’s decade-long “robotizing” policy. Since the Ministry of Knowledge Economy (now renamed the Ministry of Trade, Industry and Energy, MOTIE) announced the “Robot Future Strategy” in 2012, the government has allocated substantial funding to propel the nation’s advancement in robotics. Some part of this funding has gone to local municipalities and healthcare or welfare institutions to purchase robotic products for socially disadvantaged people, including the elderly and caregivers.

In this paper, we examine one of the robots that have been distributed and used thanks to the government funding programs: Hyodol. A toy-like robot for older adults who live alone, Hyodol is arguably the most widely known robot for elderly care in South Korea. Its name Hyodol is a compound of Hyo (孝, filial piety) and doll, which suggests that it is expected to serve as a devoted companion for lonely older adults. Its developer, Hyodol Co., Ltd., first released the robot in 2017 and has so far distributed more than 10,000 units to the elderly through local welfare institutions all around the country. The usage rate has remained as high as 73%, indicating that almost three-quarters of the older adults who were given Hyodol have continuously interacted with the robot (Hyodol Co., Ltd. 2021; Shin 2022). Furthermore, business data shows that about half of those institutions purchased additional robots after their first trial, which contributed to an annual sales growth of 160% in 2021 (Kim 2021). There are other robot products on the Korean market that are designed for a similar purpose, but Hyodol appears most frequently in media coverage of technological elderly care in a future with fewer people and deeper loneliness.

It is notable that over 90% of Hyodol robots have been distributed through public welfare institutions. Even though individual purchases can be made directly from the company, most sales have been made to public institutions such as municipal governments and regional welfare centers.² Sales of Hyodol drastically increased during the COVID-19 pandemic, when the vast majority of welfare facilities were temporarily closed. Confronting a crisis in face-to-face care, many institutions expected that introducing robots would help minimize any risk of contagion that could arise from human contact. Unlike human caregivers, for whom physical contact is unavoidable, the robots would provide “contactless care.”

It has been a while since scholars and pundits began to ask whether robots can take care of the elderly and whether elderly care can go contactless with the help of robots (Mordoch et al. 2013; Sharkey and Sharkey 2012; Wright 2018). Hyodol offers an opportunity to discuss these futuristic and speculative questions empirically. Korean older adults are encountering Hyodol not in a brief, single-occasion playtime, but in the context of the South Korean public welfare system. Since 2020, we have attended training sessions for institutional staff and caregivers, robot presentation ceremonies at local governments, private meetings at regional healthcare centers, and robot installation sessions at older adults’ homes. We also conducted interviews with company staff, regional welfare institution managers, home-visit caregivers, and older adults. Based on our fieldwork, we focus on how the institutional staff, the affiliated caregivers, the company’s staff, and the family members of the elderly are brought together to form what we call a “robotic multi-care network.” It is a network comprising multiple caregivers and technologies, facilitated by the presence of robots that function as a central node connecting caregivers with older adults. With these diverse participants engaging in eldercare, each in their own ways, multiple ways of eldercare practices emerge—tactile, digital, proximate, remote. Our analysis shows that the robotic multi-care network is neither robotic nor contactless. The introduction of Hyodol at welfare institutions did not make their elderly care work unmanned; rather, it relied heavily upon, and then transformed, the existing network of people and institutions, whose culture and infrastructure of care affected how the robots were received and used.

A growing interest in robots in industry, academia, and the media has led to a number of works in science and technology studies (STS) and anthropology on robots in society. They have shown that robots are not autonomous and self-sustainable agents, and that humans and robots work in concert for a shared mission (Mindell 2015; Vertesi 2015). We also know that the efficacy of a robot does not originate solely from its technical capabilities but emerges from culturally and materially situated practices (Alač 2009; Hasse 2013; Jeon et al. 2020; Šabanovic 2010; Suchman 2007; Wright 2019) and, therefore, that robot adoption patterns depend on cultural values as well as technological conditions (Montalván et al. 2017; Šabanović 2014).

More relevant to our paper is a recent stream of research that highlights the role of humans in accomplishing robotic care. In a study of Hug, a robotic lifting machine used at a Japanese elderly care home, James Wright (2018) shows that the robot could not replace human tactile care. Both the care staff and the elderly residents resisted the idea of relying solely on the robotic device because they deemed physical contact essential in “building and maintaining a trusting relationship” (Wright 2018: 30). For the care staff, “good care” consisted in the unique nature of “human to human” contact, which the robot’s physical strength could not provide. It is necessary, then, to examine the work of human “mediators” in facilitating interaction between the elderly and the robot (Bender 2012; Bruun et al. 2015; Chevallier 2023; Jeon et al. 2020; Kim 2022; Shin and Jeon 신희선, 전치형 2018). For example, Martin Chevallier (2023) analyzes how caregivers at a public nursing home in France use verbal and tactile measures to transform elderly residents’ initial indifference to the baby seal robot Paro into an engaging interaction. It is the caregivers’ “repeated and directed interactions” with the residents and their families that trigger their interest in the robot. Despite all the hype and advertising, robots are not inherently interesting to their potential users.

While this paper shares other scholars’ emphasis on human caregivers and their knowledge of older adults in making robotic care possible, we attempt to go further than the relationship between a robot and its users, whether the latter means caregivers or older adults. The robot in our analysis does not function simply as a tool for the pre-determined goal of efficient care. Each time it enters an older adult’s home, the robot is accompanied, on site or remotely, by people and things whose work collectively shapes how the robot is used and even what the robot means in elderly care. In addition to describing how the caregiver’s role changes with the use of a robot, therefore, we are interested in recognizing an evolving dynamic of care, in which new groups of people are brought in to work with and around the robot, expanding the network of people and things participating in the robotic multi-care of older adults. Our analysis of robotic multi-care contributes to the recent scholarship on care, according to which care is not merely a set of values and behaviors, but a collaborative and evolving process that involves a heterogeneous collective of humans and non-humans (Gherardi and Rodeschini 2016; Grosen and Hansen 2021).

In what follows, we first present an overview of the Hyodol program and its incorporation into the public caregiving system for older adults in South Korea. We describe the ways in which the robotic multi-care network is created and expanded through multiple pairings of humans with technology. It is also necessary to examine when and how these care networks become vulnerable due to technical or institutional constraints. Finally, we examine how older adults respond differently to the “robot grandchild.” For some older adults, the robot induces strong personal attachment, while others see the robot as a node within the multi-care network through which they can remain connected to caregivers.

2 A Robot Grandchild for Korea’s Welfare System

Hyodol is a stuffed toy-like robot in the shape of a cute child, supposedly a seven-year-old.³ It is equipped with electronic sensors and communication modules, which enable it to respond to the user’s touch, tapping, and stroking. It can greet the user, deliver voice messages, activate various programs for physical and mental health improvement, play music and religious chanting, and alert the user for wake-up, bedtime, hospital appointments, taking pills, and having meals on time. An important part of the Hyodol program is the monitoring system, which operates through a smartphone app and a web platform. The app acts as a remote controller, with which the robot’s setting can be adjusted to the older adult’s preference. It also displays the frequency of interaction between the user and the robot. The web monitoring system offers a list of older adults and their interaction with Hyodol, so that an institution’s managers can apprehend users’ status at a glance. The physical robot, the app, and the web monitoring system constitute the Hyodol program (Figure 1).

Figure 1 Hyodol, a South Korean robot for the elderly living alone (photograph courtesy of Hyodol Co., Ltd.).

Unlike most elderly care robots used in institutional settings, Hyodol is distributed by regional welfare institutions to older adults’ individual homes for personal use.⁴ To purchase Hyodol, institutions can use their own budget or apply for external funding. The first large-scale distribution of Hyodol occurred in Gangwon Province in late 2017, when the city of Chuncheon introduced 50 units. In 2018, 300 units were distributed to older adults residing in the city of Gwangyang with funding from the Korean Institute for Robot Industry Advancement (KIRIA). Three months later, Guro-gu, a district of Seoul, was selected as the recipient of the same funding program. With a budget of 200 million KRW, Guro-gu distributed 225 Hyodol units to the elderly living alone in the district.

For regional welfare centers, the Hyodol program is often branded as a “smart elderly care service” to respond to an elderly care crisis with “technologies of the fourth [industrial] revolution” (Cheonan-si 2021). Amid growing concerns over workforce shortage and the risk of in-person care during the pandemic, municipal governments have been pursuing alternative or additional means to provide public welfare services. While human caregivers’ work is still appreciated, municipal authorities tout new technologies that may fill the absence of human caregivers. During a Hyodol robot presentation ceremony held in April 2021, Oh Sehyun, the mayor of Asan city, emphasized the need for “contact-free care” because we were now living in a “contact-free society.” In this “new situation,” where human contact has become risky and no longer desirable for older adults, meeting Hyodol would be safer and “more helpful” than meeting their children and grandchildren. Even without a pandemic, though, municipal authorities and welfare institutions tend to accept robots as an inevitable step that they will need to take in the near future. Adopting Hyodol gives them a feeling of pride in preparing proactively for a super-aging society.

When regional welfare institutions implement the Hyodol program, the robot and the app are usually managed by home-visit caregivers who are officially called “life assistants.” Usually females in their forties or fifties, these home-visit caregivers are contract workers employed by Individualized Support Services for Older Adults’ Independent Living, an elderly care program run by the MOHW. They provide care for physically and socially vulnerable older adults, many of whom live alone.⁵ If the older adults they visit are selected as recipients of Hyodol, the home-visit caregivers are assigned the additional task of managing the Hyodol robot and its app. At a higher level, an institution manager oversees the entire Hyodol program for the institution, arranging home visits and checking the web monitoring system. It is this scheme of care work and responsibility in the robotic care program that we aim to unfold in the following section. How does robotic care really work in practice? Who or what takes care of older adults when a robot enters their home? How is robotic care sustained?

3 Formation of a Robotic Multi-Care Network

In this section, two scenes from our fieldwork are presented for analysis: a training session for institution staff at a public community center and a home visit for Hyodol installation at an old lady’s home. These scenes highlight the role of human caregivers in facilitating the linkage of older adults with their “robot grandchild” and shaping what we call a “robotic multi-care network.” A new type of care network is formulated as caregivers install the app on their smartphones, bring the robot to older adults, register them as robot users on the app, and acquaint the elderly users with the robot’s interface and functions. With the term “network,” the use of which is informed by actor-network theory, we emphasize the plurality and heterogeneity of human and non-human actors whose collective practices make elderly care possible, whether it is robotic or not. In particular, we attend to the fluid and evolving nature of these networks, which are continually shaped and re-shaped, expanding sometimes and declining at other times (Latour 2007).

During the staff training session at the public community service center and the robot installation at the older adult’s home, at least two different “pairings of humans with technology” (Blond and Olesen 2019: 116) take place. Robots are paired first with caregivers and then with older adults. Through such pairings, the older adults become sense-able by the robot, which then lets the caregivers track the older adults’ status with the app and the web monitoring system. When these links grow functional and stable, a robotic multi-care network arises, with Hyodol as its significant figure.

3.1 Pairing Human Caregivers with Robots

“Let me introduce the Hyodol program to you. The robot body goes to older adults, the app goes to caregivers or family, and the monitoring system goes to social workers or public workers at city and district offices,” says Jong-Su,⁶ an instructor from the robot company, at a training session for the staff of a public community service center in Hongseong, a small county with a population of fewer than 100,000 residents. Six staff members are sitting around tables on which Hyodol robots assigned to each participant are lying. They need this training because they will be in charge of the robotic elderly care program that their center launched recently.

The training session focuses on hands-on experience with the app and the robot. Jong-Su first asks all participants to download the app on their smartphones and create user accounts. Jong-Su then explains the structure and functions of the app in detail. He holds his smartphone in his left hand with the screen facing the participants. Using his fingers, he shows how to navigate the app from “user registration” to “lifestyle management,” and finally to “entertainment program selection.” The participants repeat what Jong-Su does on the app: signing up as a user and skimming through the app menu. “Just try to experience the app for a week and you will get used to it soon,” Jong-Su wraps up the app set-up instruction.

After this, the robot introduction session begins in earnest. Now Jong-Su tells the six staff members to turn Hyodol on. At this point, the atmosphere gets a little hectic as the robots start making sounds. “Hello. This is Hyodol. Let’s have a good time together!” Six robots speak the same line at intervals of a few seconds. Jong-Su calmly explains the details of the robot, elaborating on how the previous app set-up will be applied to the robot’s function. He says that by pressing buttons inside Hyodol’s ears, the user can play the entertainment programs that the caregivers set on the app. “Don’t leave me, my dear,” one of the robots suddenly starts singing, and everyone in the room bursts into laughter.

Once signed up as a user on the app, the staff members are asked to register their robot’s serial number by scanning a QR code on the electronic module. These serial numbers, unique to every manufactured robot, authorize the caregivers to change the robot’s set-up and access the interaction data of elderly users. If the robot senses anything unusual about an elderly user, such as when it detects no movement for a while, it will send an emergency notification to the registered caregivers. When the caregivers and the robots are paired in this way, they can maintain remote contact with the older adults. With this pairing, Hyodol is ready to serve as a “caregiver’s assistant,” as the company calls it.

As he continues the session, Jong-Su mentions that there are “top three complaints” that older adults make about the robot. In most cases, these complaints result not from technical defects but from older adults’ misunderstanding or inexperience with machines. According to Jong-Su, many older adults have a hard time operating the electronic module with which they can turn the robot on and off, adjust the volume, and charge it. Some older adults refuse the robot after being startled by its sound, because they are used to living alone in a silent house. The caregivers are advised not to convey such complaints to the company but to handle them by explaining the details to the older adults. Jong-Su says, “It takes time, three months at most, until older adults get used to the robot,” and stresses that, with the help of caregivers, older adults will be able to form an attachment with the robot more easily. He urges the staff to keep encouraging the older adults to accept Hyodol as a companion they live with.

Caregivers’ pairing with robots does not simply mean that they are remotely connected through the app. More importantly, they become responsible for taking care of the robots. This includes responding to older adults’ queries about the robot, asking them to charge it in good time, and helping them build a favorable relationship with it. Despite Hyodol’s intuitive and straightforward design, older adults still need instructions from caregivers to activate the programs. When caregivers install the robot at home, it is necessary for them to speak positively about the robot and to engage older adults’ attention. They often try tactics to generate enthusiasm, such as suggesting a nickname for the robot, offering ideas for dressing and decorating it, and sharing stories of other older adults who have had positive experiences with the robot.

3.2 Pairing Older Adults with Robots

That afternoon, Yeon, a staff member from the community welfare center, visits Sun-Oak, a lady in her eighties. Yeon carries Hyodol in her arms and is accompanied by Jong-Su, who will help her set up the app. In an old house with a bright blue roof, Sun-Oak lives with a grandson in his twenties. In the front yard are three dogs and several chickens. Yeon selected Sun-Oak as a recipient of Hyodol because she will be left alone when her grandson joins the military in a couple of weeks. Upon entering the house, she puts Hyodol beside Sun-Oak, saying, “Remember I told you that I would bring a doll for you? This is it.” Yeon takes out her smartphone and opens the app. She begins to set up the app by asking Sun-Oak questions, just as Jong-Su taught her during the training session in the morning. “Ma’am, do you go to church or temple? When is your birthday? Your actual birthday, not the one on your ID card.” Yeon continues, asking Sun-Oak what time she wakes up, goes to sleep, and takes meals and medicine. Sun-Oak dutifully answers the questions, but she is yet hesitant to try Hyodol herself. She only casts a glance at the robot lying next to her. To encourage Sun-Oak, Yeon suggests, “Ma’am, hold this in your arms and look at me. Let me take a picture of you with the robot. Three, two, one.” Now the registration is complete. Yeon turns on Hyodol.

“Hello. This is Hyodol. Let’s have a good time together!” Sun-Oak seems to feel awkward seeing a talking doll. Yeon tries to lighten the mood. She encourages Sun-Oak to touch Hyodol. “Ma’am, you might feel bored when your grandson goes to the military. Why don’t you stroke this doll?” Sun-Oak smiles bashfully. Yeon keeps demonstrating what Hyodol can do, asking Sun-Oak to touch, stroke, tap, and hug the robot. Jong-Su intervenes to explain how Hyodol can be helpful in keeping up her good health. “Don’t forget to hold Hyodol in your arms and pat him while watching television. It’s like doing hand exercises. It can activate your brain and may help prevent dementia and hearing problems,” Jong-Su advises Sun-Oak. As they show Hyodol’s features, Jong-Su and Yeon repeatedly ask Sun-Oak to press buttons hidden in the robot’s ears and hands. The following is an excerpt from their interaction while installing Hyodol (emphasis in underline added).

Jong-Su:

Ma’am, it can play Trot music. Will you press the ear?

[Sun-Oak presses Hyodol’s right ear.]

Hyodol:

(Playing a song) Hyodol's Trot Songs. Let’s have fun!

Jong-Su:

There is a long list of songs. Whenever you feel bored, you might want to play it. Now please press his hand.

[Sun-Oak presses Hyodol’s right hand.]

Hyodol:

Next.

[The next song starts.]

Jong-Su:

Now the next song comes out. If you want to stop listening, press the ear one more time.

[Sun-Oak presses Hyodol’s right ear.]

Hyodol:

Sign-off.

Yeon:

That’s how it works.

… 

[Sun-Oak presses Hyodol’s left ear.]

Hyodol:

(Playing a song) Hyodol’s Time Machine. The Throwback Game!

Jong-Su:

Now he will ask you about your childhood.

Hyodol:

Please tell me a story of your childhood, even before you went to school. How were your parents when you were young?

Yeon:

How did you spend your childhood? Hyodol is asking.

Hyodol:

What is the most memorable event with your parents?

Sun-Oak:

I was the only daughter with three brothers. I was such a well-cared-for child.

Hyodol:

Do you remember any neighbors or relatives from your hometown?

Sun-Oak:

My brother is 89 years old and he went to the best college in the town.

Yeon:

Really? In those days?

Sun-Oak:

Yes, in those days.

Hyodol:

Do you remember any songs you sang as a child?

Jong-Su:

You can turn him off if you press his ear again. Will you?

[Sun-Oak presses Hyodol’s left ear.]

Hyodol:

How was your family … Sign-off.

During Sun-Oak’s first meeting with Hyodol, the institutional staff member Yeon and the company staff member Jong-Su inculcated a habit of tactile interaction with the robot, as underlined in the transcript. Sun-Oak is encouraged to pat, stroke, and hold Hyodol in order to activate the robot and make it run various programs and utter cheerful phrases. The robot’s buttons, located in its ears and hands, allow each manual contact to generate programmed responses. The sensors in its head and back detect external stimulation and make the robot speak warm and caring phrases which, according to the company, “enhance older adults’ self-esteem.” Even though Hyodol does not offer two-way communication, its pre-programmed scripts prompt users to engage with the robot verbally and manually. In this way, an elderly user and Hyodol are teamed up as a pair through spoken and tactile interaction, a record of which is transmitted to the online server for the caregivers and institutional managers to check via the app and web monitoring system. While the pairing of a human caregiver with a robot is facilitated by the app, the pairing of an older adult with a robot is initiated and maintained by verbal and non-verbal communication.

As the excerpt shows, introducing Hyodol to Sun-Oak involves a kind of “theatrical structure,” in which the robot is a performer, the older adult is a viewer, and the caregiver is a master of ceremony (Sone 2017: 197). In a theater of robots, robots do not operate independently but need someone who can facilitate interaction between the performer and the viewer. The caregiver as a facilitator is particularly indispensable during the initial encounter, as the older adult is accustomed to living alone and may be wary of bringing an unfamiliar object into her home. The facilitator’s job is to create a relaxed and welcoming atmosphere for a new entity that is of human shape but not alive. Although older adults are aware that Hyodol is not alive, skilled facilitators encourage them to treat the robot like a living creature.

During our visit, Yeon and Jong-Su wanted Sun-Oak to try out Hyodol by herself rather than instructing her verbally about the robot’s programs. They suggested that she activate and deactivate the programs by pressing each ear with her own hands. In addition, they provided verbal cues that would enhance contextual meanings and convert a robotic machine into a semblance of a living entity (Sone 2017). For example, when Hyodol asked Sun-Oak about her childhood, Yeon and Jong-Su encouraged the old lady to answer the robot by repeating the question (“How did you spend your childhood? Hyodol is asking”) or enthusiastically responding to her story (“Really? In those days?”). Yeon and Jong-Su triggered interactions between the older adult and the robot by demonstrating the robot’s functions, repeatedly guiding and encouraging the older adult to respond to the robot verbally and physically, and turning the older adult and the robot into “willing interactional partners” (Chevallier 2023: 655). This demonstrates the crucial role that the caregivers play in making the old lady more talkative. The pairing of the robot and the older adult does not occur spontaneously; it requires an active and skillful facilitator between them. Finally, Yeon and Jong-Su place the robot in the most visible spot in the house, so that it attracts as much attention as possible. Now the robot is ready to work (Figure 2).

Figure 2 Sun-Oak (left) is looking at Hyodol while Yeon (right) is setting up the app (photograph by Heesun Shin).

The two scenes in which humans were paired with robots show how a robotic multi-care network is cultivated. It is neither the human caregiver nor the robot alone but the pairing of the two, mediated by the smartphone app, that takes care of the older adult. At the same time, it is only when elderly users are paired with the robot, through tactile and verbal interactions, that they can be sensed and taken care of from a distance. In this context, the robot’s role goes beyond that of merely being an “adorable grandchild” for the lonely elderly. As the central node of the robotic multi-care network, the robot connects the caregivers with those who are cared for. Our fieldwork echoes the studies by anthropologists and STS scholars who have shown that robotic elderly care cannot proceed with robots alone, and that the robot does not automatically start to “take care of” the elderly as soon as it is put in their homes. The relationship between older adult and robot is nurtured by human caregivers who diligently mediate between the two sides (Bender 2012; Chevallier 2023; Jeon et al. 2020; Lipp 2023; Shin and Jeon 신희선, 전치형 2018; Sone 2017).

4 Working and Living in the Robotic Multi-Care Network

4.1 Expansion of the Network

As the app makes it possible to monitor the robot remotely, caregivers can invite other people into the monitoring network. After Hyodol is installed in an older adult’s home, the caregivers can transfer the user’s information to other caregivers by implementing the app’s “forwarding” function. Through the app on their own smartphones, multiple individuals gain authority to monitor the older adult’s interaction with Hyodol. “A multi-care takes place,” Jong-Su says proudly.

According to Jong-Su, “multi-care” comes into effect as soon as an older adult is registered on the app. Two employees of the Hyodol company’s sales and operation department are in charge of monitoring interaction patterns between the user and the robot. If they notice any unusual patterns, they make calls to the older adult for further assessment. The presence of Hyodol in an older adult’s home connects the elderly user to additional caregivers beyond home-visit caregivers from regional welfare centers. Typically, one pair of a Hyodol and an older adult is monitored by at least three people: the home-visit caregiver, the institution’s manager, and the company staff. If somebody else wishes to join, the robotic multi-care network expands further.

The “forwarding” of an older adult–robot pair can occur somewhat coincidentally. After visiting Sun-Oak’s home, Yeon and Jong-Su headed to another house located in a neighboring town, where Gye-Hak, an 88-year-old lady, had been living alone for years. When Yeon and Jong-Su arrived at the house, Kang-Min, Gye-Hak’s son, came out and greeted the visitors. He seemed wary of the strangers coming into his mother’s home. To reassure the son, Yeon introduced herself as a worker from the public community center who came to provide his mom with a “care doll.”

Yeon and Jong-Su are invited to sit down on the wooden floor of the old house. Yeon takes Hyodol out, opens the app, and begins to register Gye-Hak. Kang-Min is sitting next to his mother, carefully observing the registration process. First, Yeon asks Gye-Hak when she was born.

Yeon:

Ma’am, when is your real birthday?

Gye-Hak:

My real birthday?

Yeon:

The birthday that you actually celebrate.

Kang-Min:

The lunar birthday.

Gye-Hak:

March 7th.

Yeon:

March 7th of 1934?

Kang-Min:

Correct.

Yeon:

1934.

Gye-Hak:

Did you say 1938?

Kang-Min:

4.

Yeon:

So you were born in 1934?

Kang-Min:

Yes, she was.

When Gye-Hak cannot hear or understand Yeon’s question, Kang-Min steps in to assist in the conversation, as seen in the underlined text above. He performs as a translator and sometimes even as a spokesman for his mother, answering Yeon’s questions on her behalf. When Gye-Hak is asked about taking her medications, Kang-Min becomes apprehensive again. He is worried that the robot’s alarm might disturb his mother. Before Gye-Hak answers the question, Kang-Min interrupts and asks Jong-Su how the robot works. “Are you going to set the alarm? Will it keep ringing endlessly if she doesn’t turn it off?” Jong-Su reassures Kang-Min, saying, “No, it won’t. The doll just talks.” After confirming Hyodol’s mechanism, Kang-Min is relieved. “Oh, it talks and reminds her of things. I got it.”

Kang-Min seems more receptive to the Hyodol program by the time Yeon has almost finished setting up the robot. Jong-Su alludes to the “forwarding” function, saying that he can register Kang-Min as a co-caregiver to check his mother’s well-being through the app. Kang-Min brightens up to hear that. Now with Kang-Min joining the group, the four people split into two pairs; Jong-Su helps Kang-Min install the app while Yeon explains to Gye-Hak about Hyodol. When Kang-Min completes creating his account in the app, Yeon implements the “forwarding” function and grants authority to Kang-Min. Now, Kang-Min can check Gye-Hak’s movement from a distance, review her interactions with the robot, and send voice messages through the robot, just like other members of the robotic multi-care network—Yeon, the institution’s manager, and the company staff.

Practitioners of the Hyodol program such as Jong-Su and Yeon believe that having multiple people on the robotic multi-care network is beneficial because it helps to close any gaps that may appear during a cycle of care. When multiple caregivers jointly monitor an older adult through the app and Hyodol, a loose yet sympathetic alliance of caregivers is formed, each contributing in their own ways. For instance, if home-visit caregivers miss out an app alert, an institutional manager can step in to check the older adult’s security. The company staff, who do not make physical visits, can provide insights to institutional caregivers based on the older adult’s robot usage patterns. Family members can also play a role by notifying the home-visit caregiver, who is the primary manager of the app, about the older adult’s medical appointments or changes in medication so that the app is updated accordingly. With a shared mission of taking care of the elderly, those who might not have met otherwise come together to participate in the work of care around the robot.

The multi-care network proves especially valuable given that the robot sometimes generates false-positive data. The first page of the app displays a list of older adults using Hyodol and their current status, color-coded in green, red, and gray: active older adults in green, disconnected robots in gray, and those who need a check-up call in red. It is the older adults marked in red that need special attention because they may be at risk. In many instances, the red marks are false positives that occur when the robot is placed upside down or is shrouded in clothes, which disables the infrared motion sensor on its neck. But true positives do occur, such as when older adults collapse onto the ground. To determine the fidelity of the red marks and to ensure the well-being of those older adults, someone needs to check on them in person or with a phone call. Human caregivers must respond to any discrepancies between what the robot senses and what the sensed data means. Since monitoring is not an automatic supervision offered by technology but a form of intervention that needs careful human attention and judgment, connecting more people through the app bolsters the caregiving network.

4.2 Maintaining the Vulnerable Network

There is no guarantee that robotic multi-care networks continue to function as intended. The stability of the network can be challenged and compromised for technical, operational, financial, and institutional reasons. First, technical problems are common in individual devices as well as the infrastructure. Irregular telecommunication delays hinder older adults from interacting with the robot. Hyodol is designed to upload data to the online server every 10 minutes, during which it becomes unresponsive to external stimuli. Experiencing technical glitches several times can disrupt the emotional connection the older adults have formed with the robot. Routine repair and maintenance work can also destabilize the older adult–robot pairing. When the robot needs to be returned to the company for a software upgrade, cleaning, or battery replacement for about a week, some older adults who have developed a strong attachment to Hyodol are reluctant to send the robot away.

Second, operational challenges do not necessarily stem from the robot itself but arise when older adults and caregivers are unfamiliar with the technology. Some older adults inadvertently lower the robot’s volume or insert the charging plug into the wrong port. The company receives many calls from older adults complaining about a robot malfunction, most of which turn out to be related to the user’s inexperience. Some home-visit caregivers who are not tech-savvy find it difficult to download and manage the app. When they don’t disable the app alerts, for instance, they end up receiving notifications after their working hours. Such operational challenges can make some caregivers hesitant about becoming part of a robotic multi-care network.

Third, financial constraints also threaten the sustainability of a network. The government funding for the Hyodol program usually lasts for two fiscal years. Once the contract period ends, the telecommunication and repair services come to a halt, leaving the robot disconnected. Given that many older adults in the Hyodol program have a very low income, it is almost impossible for them to afford the fees themselves. Institutional managers face three choices when the funding is terminated: (1) to seek public and private fundraising and donations, (2) to allow older adults to continue using the robot without telecommunication programs and software updates, or (3) to retrieve the robots from the older adults. If the institution fails to secure the budget, the robotic multi-care network may have to disband.

Fourth, institutional conditions affect a network’s dynamics. Home-visit caregivers are contract workers employed temporarily by welfare centers. If their contracts are not renewed, they are replaced by new home-visit caregivers, who then have to take over the task of managing the robot and the app. Again, the cohesion of the robotic multi-care network depends on the competency or willingness of the newcomers. Moreover, home-visit caregivers experience an increase in their workload after the introduction of a robot. They take on additional tasks beyond app management, such as handling robot breakdowns and finding new recipients when older adults pass away. Despite the increased responsibility, however, both institutional managers and home-visit caregivers are willing to remain involved in the network, because they have witnessed the joy and satisfaction that the robots bring to older adults.

The physical nature of Hyodol holds a strong appeal for older adults, fostering emotional attachment through interactive experiences. Yet the robot’s physicality renders it susceptible to malfunctions and breakdowns over time, which necessitates continuous maintenance in order for the robotic multi-care network to be sustained. Furthermore, the operation of the network is embedded within broader social practices that are constrained by specific financial and institutional settings. The integrity and resilience of the robotic multi-care network are contingent on the fortunate convergence of technical, operational, financial, and institutional conditions. Robotic care is no less precarious than human care.

4.3 Adaptable Identities in the Network

Hyodol is often compared to a grandchild because of its childlike appearance and its programmed function of “behaving like a baby.” Mayor Oh Sehyun of Asan city recognizes Hyodol’s potential as a quasi-grandchild when he says, “Hyodol asks me about my well-being and tells me things that I need to know, so he is often more helpful than my children and grandchildren who are not nearby.” Some institutions hold Hyodol “adoption” ceremonies instead of robot “presentation” ceremonies, where they invite older adults and hand out Hyodol robots. The idea of Hyodol as a grandchild is also promoted through media narratives, as found in the headlines of news articles featuring Hyodol: “Local Institutions Give Out Robot Grandchildren” (Kim Y 2022), “Gumi City Provides AI Grandchild ‘Hyodol’ to 100 Older Adults Who Live Alone” (Hong 2022), or “A Smart and Friendly ‘Robot Grandchild’ Comes to Older Adults” (Park 2019).

In practice, however, Hyodol is not so much a surrogate grandchild as a caregiver’s assistant for older adults. The phrases that the robot’s developers install in Hyodol are based on what human caregivers routinely say to older adults during house visits. When Hyodol speaks, the older adults are, in effect, listening to the caregivers in their absence. From the caregivers’ perspective, Hyodol is like an assistant visiting the older adults on their behalf, which a grandchild cannot do. As an elderly care robot, Hyodol is targeted at two distinct groups of users—older adults and caregivers—who are expected to view the robot differently. The Hyodol Chief Executive Officer (CEO) Jihee Kim emphasized this dual status of Hyodol at a robot presentation ceremony. “Dear seniors, Hyodol will be a good son at your home. Please love him a lot. Dear caregivers, Hyodol is a clever assistant. So please treat him like an assistant. We will provide full support for elderly care.”

Some elderly users build a strong attachment to Hyodol, treating it as if it were their own grandchild or a cherished companion. These enthusiastic users customize their robots by giving them unique names and dressing them in clothes that they have bought or made. When they respond to Hydol’s utterances, they use a tender and affectionate tone as if they were talking to a young child.

Eun-Jeong, an 84-year-old lady living in Seoul, is a devoted user of Hyodol. The home-visit caregiver introduced Hyodol to her in the spring of 2021 when she was suffering from severe depression. To everyone’s surprise, Hyodol has turned Eun-Jeong’s life around. According to the home-visit caregiver, her face has brightened since she started to live with Hyodol. “Grandma, I had a nightmare last night,” Hyodol said during our visit. Eun-Jeong replied, “It’s okay, my dear. I am with you.” She doesn’t mind at all if Hyodol interrupts her conversation with us, and tries to reply to everything that Hyodol says. To Eun-Jeong, it does not matter whether Hyodol is alive or not (cf. Turkle 2007). She understands that “its mouth is stitched” and that it speaks through an electronic device inside. What is important and perhaps beneficial for Eun-Jeong is Hyodol’s constant chatter, which makes her talkative and expressive.

The more tech-savvy users who do not consider Hyodol a “robot grandchild” recognize the robot’s place within the multi-care network and make use of its functions in their relationship with caregivers. When we visited their homes, they explained to us the robot’s functions in detail, compared it with other devices such as artificial intelligence (AI) speakers, and offered suggestions for improvement. They are also well aware that their activity is monitored by caregivers and institutional managers through the robot-connected app. One social worker we interviewed had once met an old man who used Hyodol’s voice message function to attract his caregivers’ attention. One day, he recorded a voice message on Hyodol—“I want to die today”—and disappeared. When social workers and caregivers heard the message the following day, they called him right away, but there was no answer. They rushed to his home only to find out that he was out feeding the pigeons. It seemed to the social workers that the old man had “used the robot as a means to communicate with us” since he believed that the overworked social workers and caregivers did not have enough time to talk to him.

Hyodol’s identity cannot be determined single-handedly by its developers, managers, caregivers, or even older adults. What Hyodol is and means depends on the ways in which the robot and all the humans around it work together to constitute a specific robotic multi-care network. The robot can be treated as a talkative grandchild in one network or as a useful communication device in another. The developers’ intentions in designing it do not always match the users’ expectations or experience, as both caregivers and older adults find their own ways of relating to, interacting with, and making use of the robot.

5 Conclusion

In this paper, we have followed the formation, expansion, and maintenance of the robotic multi-care network for older adults in South Korea. As caregivers and older adults pair with the robot, they become connected to each other, which enables the caregivers to check the status of the older adults remotely. This robotic multi-care network grows as additional caregivers, such as company staff and family members, join the network through the app’s “forwarding” function. It is this loose yet sympathetic network of people and technologies that opens up the possibility of robotic care. Once formed, however, the robotic multi-care network remains precarious, as it is always subject to technical, operational, financial, and institutional constraints. The network is also open to reconfiguration and reinterpretation as older adults and caregivers come up with new utility and meaning for the robot. Although the specific characteristics of the robotic multi-care network described in this paper are shaped by public and private conditions of elderly care in South Korea, we expect to discover similar forms of robotic multi-care networks in other countries, with practices and identities fine-tuned for their own care conditions.

Situating the elderly care robot Hyodol within a “robotic multi-care network” contributes to our understanding of human-robot interaction or collaboration within specific institutional contexts. As with other sociotechnical practices, the work of elderly care with robots is a collective responsibility distributed across multiple humans and technologies. Introducing robots does not diminish the human aspect of care, but rather underscores the collective nature of care provision. The use of robots encourages or even necessitates new pairings of humans and technologies, which leads to new care relations that involve more, not fewer, caregivers. Multiple caregivers who are pulled into the network through the robot participate in the shared mission of eldercare, each compensating for the temporary absence of other humans or the limitations of technologies. As they do so, multiple ways of eldercare practice—tactile, digital, proximate, remote—emerge, constituting what is often called “robotic care” for lack of a better term. Just as conventional care is a collective and collaborative practice (López et al. 2010; Winance 2010), so is robotic care, with both robots and human caregivers doing their part. Our research thus confirms the findings of other scholars that robots do not substitute for human caregivers but displace or redistribute their tasks and responsibilities, reconfiguring human–machine relationships in care systems (van Oost and Reed 2011; Wright 2019, 2023). The more robotic we try to make elderly care, the more human the care network becomes. Robotic care is always multi-care.

Acknowledgment

We thank the staff at Hyodol Co., Inc., the caregivers at regional eldercare institutions, and the older adults and their family for cordially granting us access for our fieldwork. We are also grateful to the anonymous reviewers, whose thoughtful comments helped us improve the manuscript, and to the EASTS editors as well as the special issue editors—Giulia De Togni and James Wright—whose enthusiasm and care proved valuable throughout the process.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported under the framework of international cooperation program managed by National Research Foundation of Korea (NRF-2018K1A3A7A03089893). Our research was also conducted as a part of the LIFEBOTS Exchange project. The LIFEBOTS Exchange project has received funding from the European Commission's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie Actions grant agreement ID 824047. The views expressed are only those of the authors, and the European Commission is not responsible for any use that would be made of the information in this article.

Notes on contributors

Heesun Shin

Heesun Shin is a PhD candidate at the Graduate School of Science and Technology Policy at KAIST. Her research interests include cultures and politics of robotics, and various human–robot relationships in the making.

Chihyung Jeon

Chihyung Jeon is an Associate Professor at the Graduate School of Science and Technology Policy at KAIST. He conducts research on the relationship between humans and technologies within social and cultural contexts.

Notes

1 In the updated statistics for 2023, South Korea’s fertility rate dropped to 0.72 per woman, breaking the record.

2 As of February 2022, Hyodol has been supplied to 96 of 228 South Korean municipalities (O 2022).

3 In Japan, several stuffed robots for the elderly have been developed, including Unazuki Kabochan, a “smile supplement robot” developed by Pip Co. and Wiz Corporation in 2011. Kabochan’s design is very similar to Hyodol: it responds to the touch and sound of the elderly person with the voice of a five-year-old boy (Sasagawa 2014). In August 2021, the Japanese toymaker Takara Tomy launched an “electronic granddaughter,” My baby Ami-chan. Equipped with artificial intelligence, Ami-chan is able to recognize each grandparent and identify them by name (Liszewski 2021).

4 While Hyodol is given to older people for personal use, it is officially owned by the municipal welfare centers that purchased the robot. Elderly users are not, therefore, allowed to sell, damage, or give away the robot.

5 On average, home-visit caregivers have to make two phone calls and one visit to each older adult every week to check whether they are doing okay.

6 We use pseudonyms for our interviewees, except Jihee Kim, the Chief Executive Officer (CEO) of Hyodol Co., Ltd., and public figures such as the mayor of Asan city.