Human communication has been studied for many years. It involves both verbal and nonverbal language. Nonverbal language includes touch, gaze, and gestures which humans use unconsciously when communicating. As machines become more advanced, they begin to respond to our verbal commands and are able to move similar to humans. Despite this advancement, there are limitations and boundaries that will need to be put into place.
The first feature to focus on would be touch. Touching someone is an intimate act that can induce physiological, emotional, and behavioral responses (Willemse, Toet, & Erp, 2017). Whenever someone is sad, the physical action of a hug causes the person to have an emotional response. According to Willemse, Toet, & Erp (2017), a simulated touch may replicate these responses. Willemse, Toet, & Erp (2017) discovered that touch by a robot did not reduce stress responses or deflect the participants attention from the stressor, the participants’ perception of the robot did not change, and a touch did not increase or decrease the willingness to donate money. Not everyone is receptive to a touch as well; there would have to be boundaries for how often touch is applied.
A second feature to focus on would be human language. Human language is not precise. This could be an issue with human-robot interaction especially since human-friendly robots will be used by non-experts and will require the use of more human-like interactions (Muthugala & Jayasekara, 2018). According to Muthugala & Jayasekara (2018), early human-robot interaction (HRI) was only able to follow simple, single-word commands. This is a limitation that needs to be lifted especially with a care robot. There is also the issue of distance as measured in human terms. Muthugala & Jayasekara (2018) also showed that environment plays a part in distance. For example, when we ask someone to place a teapot close to a candle, we do not specify how close to the candle we want the teapot. Muthugala & Jayasekara (2018) indicates that this is imprecise language. The teapot in the example could be touching the candle or could just be occupying the same pane of glass on the table. There was no exact distance specified which a robot would require. Obstacles in the room that the robot is in must also be considered (Muthugala & Jayasekara, 2018). Another example is having a coffee table in front of the couch and asking a robot to move to the window on the opposite wall; the robot will run into the coffee table while humans would walk around it. If the coffee table was removed, the robot would be able to go towards the opposite wall. Robots would require sensors otherwise. Bao, Jia, Cheng, Tang, & Xi (2016) used a camera and included color, shape, and material into their robot’s database. Humans use these descriptions unconsciously when communicating.
A third feature to focus on is that human communication is not completely verbal; there are people who communicate in various other ways such as American Sign Language and written communication. Those are two nonverbal forms of communication. Care robots would need to be equipped with cameras or linked to a tablet that would allow the care robot to communicate with those nonverbal individuals.
A final factor from the human side is ethics. What may be considered ethical in one culture, may not be ethical in another. There is one component that is consistent across all cultures and that is the basis of the Hippocratic Oath; that is for healthcare professionals (and especially doctors) to do no harm. According to Sio & Wynsberghe (2015), humans should “not just [analyze] the impact of care robots on the basis of one prominent ethical consideration or approach.” This means that care robots must be programmed to perform ethically and not as a retroactive approach (Sio & Wynsberghe, 2015). Also, Sio & Wynsberghe (2015) states that there is no appearance standard for a care robot, but there are four elements that are part of the fundamental care values: attentiveness, responsibility, competence, and reciprocity. Humans require other humans to form a basic form of companionship. Care robots, if predominant, would decrease real compassion of human companionship (Sio & Wynsberghe, 2015). In addition, some elders may not feel comfortable with a care robot and would prefer an actual human. It was also determined that there should be limits even with medically competent elders when a robot assists with a transfer such as making sure the care robot does not lift them over a balcony edge except in an emergency (Sio & Wynsberghe, 2015).
To conclude, the limitations that currently exist are poor understanding on the robot’s part of natural language both verbal and nonverbal and the ethics involved in using any sort of robot. The feature that need focused on the most is the natural language communication between humans and robots and how it may be accomplished with those whose primary mode is nonverbal. Ethics must also be considered and programmed to prevent potential issues when a human user asks the robot to do something that another human would immediately question and refuse to do.