(1) Autonomous robots [9] are intelligent machines capable of performing tasks in the world by themselves, without explicit human control. Examples range from autonomous helicopters to Roomba®, the robot vacuum cleaner.

(2) Healthcare robots (HRs) [10] are these robot used in the healthcare setting. A Healthcare robot can be used for healthcare setting, but it does not retain the adequate competency of “caring in nursing” like a human nurse.

(3) Humanoid nursing robots (HNRs) assumes them to be a fully functional robotic nurse who can do the same job as a human nurse as addressed in the next section on true and genuine nursing care for human beings.

Two theoretical perspectives underpin the structuring of the recommendations towards the design and development of HNRs. These theories of Nursing provide the essential elements, characteristic ingredients, and attributes of HNRs considering the discipline and practice of nursing that is described within the founded expectation of performance levels of nursing care.

Watson (2010, 2012) [11] [12] defines nursing as “a human science of persons and human health-illness experiences that are mediated by professional, personal, scientific, esthetic and ethical human transactions.” Watson’s theoretical underpinnings have seven assumptions: Caring can be effectively demonstrated and practiced only interpersonally. Caring consists of caritas factors that result in the satisfaction of certain human needs. Effective caring promotes health and individual or family growth. Caring responses accept the patient as he or she is now, as well as what he or she may become. A caring environment is one that offers the development of potential while allowing the patient to choose the best action for him or herself at a given point in time.

In contrast, Locsin (2005) [13] acknowledges the practice process of nursing as “knowing persons as caring”. When nurses care for patients, nurses strive to understand the patient as a whole person. Within this theory, technology is used to continuously apprehend and express persons as caring who are ever-changing and unpredictable. Nursing is a knowledgeable practice provided within a human-to-human relationship in which caring is expressed in a continuous pro- cess. In order to understand persons, it is essential to enter into their world, recognize, understand, and appreciate who and what they are as whole persons. It is therefore necessary for high quality nursing practice that nurses are trusted by their patients and for patients to be empathically understood as human beings in a technology-dependent world of healthcare.

When “knowledge, judgment power, technical ability, and caring competence” are required in a particular situation, the nurse should be able to demonstrate “technological competency as caring in nursing.” In demonstrating this competency, the nurse is envisioned to practice nursing using the process of knowing persons, in which technological knowing, mutual designing, and participative engaging [14] are the shared events. For example, in a critical situation, competent nurses can observe patients who may be suffering from severe headache, and who cannot sit and stand. The technologically competent nurse enters the patients’ world of suffering and can be quick to assist such patients. On the contrary, a nurse who cannot distinguish patients’ condition is unable to know the person fully in the moment, thereby fail to address the patient situation. As such, these nurses incompetent at that moment. The patient in the critical situation is seeking support and affirmation from nurses and other health care personnel; therefore, if the nurse does not notice the call (silent cry) for nursing, this nurse has failed to affirm the wholeness of the patient. Such a behavioral illustration is much like that of an office worker whose education is not of professional nursing. If and when the nurse does not understand and empathize (sympathy makes the nurse the authority of the pain instead of the patient who is experiencing the pain) with the patient’s pain, and simply measures the blood pressure/body temperature and perhaps only listen to the patient’s condition, such tasks can be possibly performed by reception computers as early as when the patient is first encountered.

It is but clear that what the patient is seeking is a nurse who expresses technological competency as caring, who will know him/her as person with hopes, dreams, and aspirations to be who he is, and wants to be [15] and be affirmed, supported, and celebrated as being a person. Nevertheless, the expression of true and genuine nursing is always changing as persons also continuously change.

Caring represents an essential human need and is the fundamental constituent of professional nursing practice.

Mayeroff (1977) [16] in his book “On Caring”, recognized and described eight ingredients that comprised “caring”, namely, knowing, alternating rhythms, trust, hope, courage, honesty, patience, and humility. Likewise, Roach (1987) [17] has determined six attributes of caring, namely, compassion, competence, commitment, conscience, confidence and comportment which comprise the viewpoint of persons as caring. These ingredients and attributes constitute the characteristics of a caring person guided by Boykin and Schoenhofer’s (2001) [15] theory of “Nursing as Caring.” Nursing practice is described as the shared lived experience of the caring between the nurse and the person being nursed. In the process of caring, nurses develop and understand the knowledge and skill in caring constituting the knowledge base of nursing [18] .

A deeper appreciation and understanding of caring and its expressions in nursing helps nurses understand other persons thereby influencing how they provide high-quality care [19] . The act of nursing is an excellent intervention performed by highly competent nurses, in which advanced observational abilities for patient situations, e.g. those with dementia, and intentional communication skills for patients are necessary.

It is difficult to determine and rationalize whether or not it was the prevailing culture of the day that implicitly prioritized technology and cure over caring and non-technological interventions that made nurses feel undervalued and virtually invisible within the decision making process in most health care facilities [20] . Health and health-related actions do not occur in isolation but are related to one another in time, space, and content [21] . Nurse caring was reported as an “interactive process that occurs during moments of shared vulnerability between nurse and patient” [22] . Processes of nursing practice [23] comprise the activities of caring, these include nurse attributes for professional caring (structure) which can be viewed as nurse interventions, and the dual nature of caring which encompasses attitudes and actions (process).

The dimension of interest is the opposite in that empathic capacities that robotics and artificial intelligence (AI) can demonstrate for humans can exist through programmed activities. Regardless, the knowledge generated will bring information to engage in relationships between empathy and AI and contribute to understanding its usefulness and impacting nursing/caring theories [24] . When the HNRs are able to express themselves through emotive behaviors much like humans can they will be able to convey empathic understanding [25] in which the benefactor will be the patient.

Figure 1 describes the expected relationships between and among nurses, HNR, and patients. The ability of HNRs to recall/respond to critical situations in patient care management is crucial. In situations such as these, it is important to share the experiences of patients. The functionality of HNRs to process language and maintain a conversation and to read and convey emotions are integral to the success of HNR relationships with and among nurses and patients.

Hubert Dreyfus and Stuart Dreyfus’ work [26] was extremely influential in Benner’s theory development. Adapting this model to nursing, Benner outlined the development of skill and learning in clinical nursing practice in her theory “Novice to Expert” [27] . Dreyfus and Dreyfus’s novice to expert model can be found in Hubert Dreyfus’s book What Computers Can’t Do [26] , first published in 1972. Hubert Dreyfus was a professor of philosophy who argued that the so-called AI displayed by computers was, in fact, nothing of the kind. Computers, he argued, can be programmed to make logical decisions based on an extremely rapid analysis of huge amounts of information, but can never be programmed to be intuitive or even to display common sense.

Benner [28] [29] detailed the acquisition of nursing expertise and proposed five possible expertise levels: novice, advanced beginner, competent, proficient, and expert. Nurses at the novice stage are still in nursing school. Especially, competent nurses are task-oriented and deliberately structure their work in terms of plans for goal achievement. Proficient nurses perceive situations as a whole and have more ability to recognize and respond to changing circumstances.

Experts have an intuitive grasp of whole situations and are able to accurately diagnose and respond without wasteful consideration of ineffective possibilities. Thus, it may be useful to develop AI like the growth and development of human nurses.

Current development of AI is addressing issues of creating AIs to be more intelligent than humans [30] , and ensuring that they use their advanced intelligence for good rather than ill effects. Due to the exponential evolution of future AI, the relationship between human beings and machines can profoundly change rapidly. There is that possibility that AI can actively learn.

Several studies exist that currently provide information on the topic of HRs’ functionalities [25] [31] [32] [33] [34] [35] . Findings of these studies direct the perfection of “real” HRs in the future, those that can assist with the demands of human caring in nursing.

HRs that can perform simple tasks like taking vital signs or perform precise delivery [36] , for example medications are critically needed. Reducing these types of human-dependent activities will provide human nurses the more important focus on direct human caring relationships and nursing care for these patients. Furthermore, HRs research [37] is urgently needed to focus on developing new computational algorithms for determining accurate patient emotional state classification in interactions between human and intelligent machine relations with HRs during health care services.

Any HNRs design to “replace” or enhance the work of the nurse and increase patient satisfaction must be cognizant of this relationality [38] . In this way, it behoove researchers of Computer science, Robotics, and Engineering to muster and obtain increased appreciation of the value of nursing care practices in health care, realize hints from the information and recommendations provided, and be able to design and create a new robot technology to support nursing practice. Therefore, in order to make the humanoid robot perform nursing, it requires advanced performance of real time nursing and the ideal HNRs need to have self-aware ability.

In order for HRs to interact with older people through a spoken dialogue, a high performance speech recognition system for speech recognition of elderly people is needed. It is necessary to develop the new speech data corpus to improve speech recognition accuracy of the speech of older people. This is particularly critical as interaction with such robots is expected to help among other aspects, the maintenance of the cognitive abilities of nursing home residents, while also providing them with companionship [39] .

Figure 2 illustrates the relationship expected of HNRs, e.g. among the elderly, regarding its performance as based on its type and levels of function. Requisite nurse-robot activities are especially anticipated in assessing patients who need care and the characteristic performance of HNRs within the human-robot “caring” relationship. This is often expressed in the human-robot interface that is critical to expressions human caring that ought to be much like those conveyed by human nurses.

The increasing population of older persons world-wide fosters the recognition and increasing realization of the need of a theory-based practice model that can guide professional nursing practice in situations in which HRs are engaged with human persons (patients). Being critical participants (patients) in the process of human health care, transactive engagements of persons with HRs has become increasingly possible to envision, in the next few decades. A theory-guided nursing practice is integral to meeting the demands of this vulnerable population. Attaining quality health particularly in situations in which human beings, robots, and patients are involved may be the common health care practice in the very near future.

The realization of human and intelligent machine transactions in human health care has been the impetus for the development of the Transactive Relationship Theory of Nursing (TRETON) [40] . As the demands and needs for quality health care rise, particularly in human-resource poor environments and among the older adult population, the utilization of human-machine process requirements are intensified, and the realities and consequences of transactive relationships have become integral to assuring quality human health care.

Nursing caring practices involving intelligent machines with highly sophisticated robotic technological functions such as interactive capabilities may be the best option for the care of patients with the increasingly demanding human care such as the older adult population. In this situation, the HNRs with AI will be programmed to appreciate the meanings of the lived experiences of persons. Therefore, within the Transactive Relationship Theory of Nursing (TRETON), a nursing encounter where all nursing occurs encompasses the process of nursing as technological engagement and mutual knowing.

Discovering and deliberating the situation about AI and anthropomorphic machines in nursing and health care, before these become ordinary is of great importance. Erikson & Salzmann-Erikson [25] [41] have mentioned that it is highly likely that AI will be implemented in nursing robotics in various forms, both in medical and surgical robotic instruments, and also as different types of automatons/droids and humanoids.

The basic question about HRs and HNRs and its development is, “should the aim of nursing be focused on developing humanoid nurse robot transcending AI and surpassing human intelligence and physical function?”

One of the popular robot functions in nursing today focuses on functionalities of nurse robots-the so-called “nursebot.” In this robotic assistants project [42] , the goal of the robot software system was to equip robots with intelligence programming so that these can make ethicomoral decisions, for example, about what reminders to issue and when, and balancing potentially competing objectives such as: (i) ensure that the user is aware of activities she/he is expected to perform, (ii) increase the likelihood that she/he will perform at least the required activities (e.g. taking medicine), (iii) avoid annoying the user, and (iv) avoid making the user overly reliant on the system.

During the course of experiments conducted in an assisted living facility, the robot successfully demonstrated that it could autonomously provide reminders and guidance for elderly residents.

Figure 3 illustrates the expected functionalities of intelligent machines identified and described in four levels and the projected timeframe regarding its advancing functionalities (Please see Section 3.4).

Ÿ Level 1: HRs in 2020, are low level “bots” that works much like a HNR does. This intelligent machine will do vital sign measurements: blood pressure, pulse measurement, percutaneous oxygen saturation measurement, clerking, and so on. It is this public image of the nurse particularly in hospital settings that provides a semblance of its ontological and epistemological work. Activities such as these can include bathing, assisting with meals, taking vital signs, making beds, assisting physicians etc.

Ÿ Level 2: HRs in 2030. This type of robot can manage and perform moderate level of work. HNR is a robot with interpersonal communication, in addition to the expected work of low level nurses, can understand, empathize, and share empathy with patients and families. This level of robot nurse can also interact with other professionals. It may seem inconceivable that this so called “excellent nursing” practice is the outcome of competency of practice in human engagement involving physico-physiological and interactive relations with and among other persons as patients. If we are to make the nurse robot's heart/mind necessary for nursing equipped with AI, it needs the ability to first observe patient's need, correctly evaluate it, and communicate its findings to the patient in appropriate words. Also nurse robots communicate their empathy with technological competence. For example, a series of actions that integrate not only words but also “knowledge, judgment, technical skills, and care” are expected of these nursing robots.

Ÿ Level 3: HNRs in 2040, will be equipped and expected to perform high level of work as HNR assumes to be a fully functional robotic nurse who can do the same job as a human nurse. It is important to be reminded that all levels of robot nurses are equipped with terminals of electronic medical charts for easy access.

Ÿ Level 4: HNRs in 2050, by this time, transcendence becomes the feature of robot functionality. It refers to the occasion when AI exceeds human intelligence and physical functioning. Should we aim to develop the HNR towards transcendence?

Usually the singularity [43] a hypothetical moment in time when AI and other technologies have become so advanced that humanity undergoes a dramatic and irreversible change “maybe a hoped-for occasion.”

Many definitions exist, but one that serves AI and humanity well is what Kurzweil [44] has described, that singularity [44] is a trait marking one as distinct from others; a peculiarity. Often, computers manage to illustrate a hypothetical future point in time when AI will surpass human intelligence and be able to self-replicate and improve itself autonomously. If it is something uncommon or unusual, it should be a HNR enriching human society.

While Singularity is a conditional “reality” for HNRs’ existence, the question of machines potentially creating disorder to humanity cannot be simply ignored. Ito, et al. [45] has addressed this viewpoint studying concerns of present-day health care workers about practicing their professions with these HNRs. Importantly, we must not forget that HNR enriches human life. In other words, we should not create things that threaten our lives.