CHAPTER 7

Human-Robot Interactions: Will You Become BFFs with a Robot?

Computers will be at human levels, such as you can have a human relationship with them, 15 years from now. …When I say at human levels, I’m talking about emotional intelligence. The ability to tell a joke, to be funny, to be romantic, to be loving, to be sexy. That is the cutting edge of human intelligence.

—Ray Kurzweil, Director of Engineering at Google, inventor (inducted into the US National Inventers Hall of Fame) and technology author (Thompson 2014)

Chapter Overview

In this chapter, we explore a fundamental question: will humans ever form truly meaningful relationships with robots? Based on an exhaustive review of past research, we come to the conclusion that yes, after some time to adjust and overcome the initial awkwardness, many humans will form meaningful, personal bonds with robots, though certain conditions are required. The necessary conditions include robots conveying “social presence,” meaning being perceived as truly part of an interaction via their intelligence, speech, and physical gestures (though no actual robot self-awareness or consciousness is needed). It also involves robots conveying emotional intelligence, by reading and responding to our emotions as well as expressing emotions of their own. And, finally, to create human-robot bonds, robots will need to express their own individual personalities. We discuss all these conditions in-depth, and review how robot developers are currently working on them. We then talk about three hierarchical categories of human-robot relationships: (1) robots treated as tools (minimal relationship), (2) robots treated as pets (moderate relationship) and (3) robots treated as humans (most significant relationship), with variations within. We end the chapter by discussing the implications of each relationship type.

Initial Awkwardness When First Conversing with Robots

Expectations are that robots will increasingly become an integral part of our lives - in our homes as well as our communities. They will be up close and personal with us. How will we humans relate to them? Will our interactions with robots be cold and functional? Will we view them as smart tools and nothing more, similar to how we view our current computers? Or, will we actually bond with robots? Will we have feelings of friendship, caring, and even love toward them, as we do toward many humans—and animals—in our lives? Research conducted over the past several years suggests that we will indeed bond with robots, but it will take a while and the robots will have to achieve certain functionality before it can happen on any widespread scale.

When first interacting with social robots, meaning intelligent robots we are meant to talk and deal with interpersonally, there will be a high degree of awkwardness. It will be uncomfortable to treat something we know is not alive as if it is. Research suggests that when people are first interacting with these more advanced social robots they will feel: (a) uncertain, not knowing the best way to conduct the interaction and (b) less enjoyment, it will be less fulfilling than an interaction with a human (Edwards et al. 2016; Spence et al. 2014). This initial awkwardness will apply to situations such as: a store patron interacting with a customer service robot while the robot helps the patron find a new outfit for an upcoming wedding; a hotel guest asking a robot concierge about the best restaurants in town, or an elderly resident of an assisted living facility talking about the day’s football games with one of the institution’s companion robots.

This awkwardness is due, in part, to the issue of perceived social presence. Social presence is a concept used by academics who study interpersonal communications, and it is defined as being “real and present” in the interaction, without any barriers or artificiality in the way. Two humans talking with each other in person, where both individuals are highly focused on and engaged in the conversation, constitutes high social presence. Two humans having a conversation with each other via video chat involves less social presence, since not all the communication cues (such as hand gestures) can be seen and the technology limits, to at least some degree, a full interaction when compared to an in-person situation. Two humans talking on a telephone in audio-only mode reduces the sense of social presence even more. Social presence is a concept that has been associated with human interactions, and how technology can help or limit it. Recently, however, it has been applied to human interactions with robots, where humans are talking with technology rather than via technology. This research suggests that people will perceive interacting with social robots as involving less social presence. Even though the robot will be talking, moving, and reacting to our comments, we humans will still see the robot as being less “real and present” in the interaction, simply because we know robots are artificial devices (Edwards et al. 2016; Spence et al. 2014).

This sense of less social presence in the interaction will lead to greater uncertainty. People will be unsure how to interact with a robot, at least at first. Should we follow all the typical norms of interactions, such as politeness, when talking with robots? For instance, does it make sense to say “please” and “thank you” to a robot? After all, we do not thank our computer when it performs a function for us. But we would thank a human concierge if she recommended a fantastic local restaurant to us. Also, this sense of reduced social presence will lead to less enjoyment. It simply will not be fulfilling and satisfying to humans to interact with something that is seen as not truly alive and not truly present in the situation.

This will be the reaction at least initially. People often feel uncertain and awkward doing anything that is new, yet such feelings often dissipate with repeated experiences. Research suggests this will also be the case for human-robot interactions. Researchers in Japan conducted a study that included having study respondents repeatedly interact with a social robot. Measurements, which included post-interaction questions asked of the humans as well as their body language during the interaction, showed that they indeed started off feeling uncertain and experiencing less enjoyment, but after repeated encounters the interactions with the robots began to feel more normal to them (Haring, Matsumoto, and Watanabe 2013). This suggests that as we humans get used to robots in our lives, and if the robots are responding to us in a manner that is humanlike, we will begin to anthropomorphize the robots, feel more natural, and get more comfortable interacting with them and even begin to bond with them. The initial awkwardness will wear off.

Anthropomorphizing Robots

Anthropomorphism is the attribution of human traits, feelings, and intentions to nonhuman entities. This appears to be a natural tendency for humans, as we do this regarding animals (especially our pets) as well as inanimate objects such as our cars (Hutson 2012). There is significant evidence that humans will anthropomorphize robots, meaning treat them as if they are living and breathing entities, even if we know they are not (Zlotowski et al. 2015). Consider this true story:

In 2015, robotics company Boston Dynamics released a video clip introducing ‘Spot,’ a distinctly dog-like robot. In the clip, Spot is kicked twice by humans and scrambles to stay hard on all four legs. The purpose of kicking Spot was to demonstrate the robot’s stability, but many commenters took to the internet to express discomfort and even dismay over Spot’s treatment. The slew of negative reactions even compelled animal rights organization PETA to acknowledge the incident. (Darling 2017)

We have to assume the individuals viewing the video knew Spot was a robot, as the robot was very mechanical looking and the video clearly stated it was a robot. And yet, viewers expressed pain and outrage seeing it kicked. This is not the only case of an emotional reaction to a robot being hurt. There are reports of US soldiers expressing emotional distress over the destruction of robots they worked with on the battlefield to help locate mines, even though the soldiers know these robots could be quickly replaced by other robots (Garreau 2007). This suggests these soldiers are thinking of these robots as comrades in arms, at least to some degree similar to their fellow human soldiers. If soldiers can react in this manner, clearly anyone can.

The tendency to anthropomorphize robots will be driven by many factors, particularly the humanlike features given to them, their apparent self-direction and their intelligence (Riek et al. 2009; Epley, Waytz, and Cacioppo 2007). Robot designers have been and will continue to emphasize features that are meant to make us even more likely to anthropomorphize robots. The goal is to make relationship building with robots more natural. After all, robot manufacturers want us to bond with their creations.

Bonding with Robots

Humans are social animals. We have evolved that way over hundreds of thousands of years because by living and cooperating in cohesive social groups we could better protect ourselves from predators, better obtain food, and better care for our offspring. Hence, building social relationships is hardwired into us as social beings. Most of us want to bond with others around us. Bonding is defined, rather simply, as “the formation of a close relationship” (merriam-webster.com). We would add that successful long-term bonding involves feelings of affection and affinity.

This extends beyond human-to-human bonding. Humans regularly form deep bonds with animals, even though we understand that animals think and perceive the world differently from us and have their own peculiar instincts which drive much of their behavior. Yet, for tens of thousands of years we have established deep and meaningful relationships with animals. This is particularly true of dogs, which have been “man’s best friend” for tens of thousands of years (no offense to all the cat lovers out there).

Researchers have examined the long-term relationships that have flourished between humans and dogs. They have found these relationships often surpass a simple pet relationship and take on the form of a bond more akin to friendship, much as between two humans. Studies have found that the success of human-dog bonding is based on many factors. First, emotions are a key ingredient in bonding. Most humans cannot bond with an entity that lacks the ability to read human emotion or display emotions of its own. Dogs can do both. Second, a unique personality needs to be expressed by both parties for bonding to occur. Humans need to believe they are interacting with an individual entity with its own uniqueness. Dogs have this. Finally, there must be benefits to both sides for a bond to work—some sort of positive feedback or satisfaction (Konok et al. 2018). The example of human-dog bonding gives us some fundamental ideas to think about regarding possible human-robot bonding.

Academics who study robots believe humans and intelligent robots can form meaningful bonds (Edwards et al. 2016; Graaf 2016; Turkle 2010). By meaningful bonds, these researchers mean humans can develop feelings of affinity, friendship, and caring for intelligent robots—much as is the case between humans and animals and even humans and other humans. These bonds will not happen right away and they will not happen in every case. Based on previous research and our own studies with consumers, the following factors need to be in place for meaningful human-robot bonding to occur:

1. Social presence (on the part of the robot, as perceived by humans)
2. Emotional intelligence (on the part of the robot, as perceived by humans)
3. Individual personality (on the part of the robot, as perceived by humans)
4. Personal benefit (for humans)

We will now explore each of these in greater detail, and discuss what robot manufacturers can do to help strengthen the likelihood that humans will want to bond with their robots.

Bonding Factor #1: Social Presence

Social presence, for human-to-human interactions, is defined as the degree of salience of an individual in an interaction, meaning to what degree the individual is perceived of as “real and present” in the interaction by his/her counterpart. When we interact with someone else, we do not want to feel as if we are talking to a wall. We want to feel as if we are talking to a person who is there, present in the moment and truly interacting with us—with no barriers and no distractions. The same is true of human-robot interactions. There are three characteristics of robots that can help bolster a human’s perception of its social presence in an interaction: intelligence, appearance, and physical gestures.

Intelligence. For a robot to be perceived of as truly socially present in an interaction, the robot needs to be intelligent enough to carry out a meaningful conversation with its human counterpart. It must respond quickly and naturally to questions and comments, and maintain its part of a dialog with no pausing, no confusion, and no other perceived barriers or distractions. Artificial intelligence and machine learning are getting closer and closer to the ability to pass the Turing Test. The Turing Test, proposed by the famed computer scientist Alan Turing in 1950, is a test of a computer’s ability to exhibit intelligent behavior indistinguishable from a human. When this happens, when a robot can converse with us as normally as a human, we will be less likely to think we are talking to a pile of metal and plastic and instead believe we are talking to an entity that is truly socially present.

Appearance. As discussed in Chapter 6, a robot’s appearance is critical to how a human will perceive it. A humanoid robot appearance is the optimal approach for promoting natural human-robot bonding. This means the robot would have a humanlike face and body, though without attempting to look exactly like a person (which, unless done perfectly, is perceived as frightening). A humanlike face is particularly important, with large expressive eyes, since humans naturally look into people’s eyes when conversing (as Shakespeare told us, “The eyes are the window to your soul.”) Also, a mouth from which, or near which, the robot’s voice emanates will make communications more realistic. A more humanlike, but not frightening, face for the robot (which can be fixed or on a screen) will aid in a human counterpart perceiving the robot as being socially present in the interaction. Consider the faces of numerous robots already featured in this book, such as Hilton Hotel’s Connie the Concierge (Chapters 1 and 6), Sanbot’s Elf and Mabu from Catalia Health (from Chapter 5) and NAO from Alderan robotics (Chapter 6) and Buddy from Blue Frog Robotics (Chapters 5 and 6).

Gesturing. Human conversations involve more than simply talking. Physical gestures are a critical part of natural human communication, which involves how we move our heads, our hands, arms, body, and eyes. Researchers suggest that robots must master the complex array of human gestures used in communication, which of course vary between cultures. Mastering physical gestures will aid in humans seeing robots as humanlike participants in social conversations (Graaf 2016; Sidner et al. 2005). Hence, the robots viewed as most socially present will use appropriate eye contact, looking into the eyes of their human companions appropriately enough to show interest, but not so much to appear threatening or unnatural. These robots would also use hand and arm gestures properly for emphasis. And these robots would nod, tilt, and turn their heads as a human would during social interactions. The more natural a robot appears during its interactions with a human counterpart, the more the human will anthropomorphize it and believe it is socially present.

Is consciousness necessary? A significant question you may be asking as you read this is: does a human need to believe a robot is conscious, meaning self-aware, for the human to believe the robot is socially present? Certainly, this would greatly help. In such a case, the human would see himself/herself as talking to an entity that is truly alive. However, consciousness is a complex concept, and there is currently significant debate as to whether or not robots or any technology can ever truly become conscious. Even if a robot has not achieved consciousness, but is doing all the things mentioned earlier – meaning holding its end of a conversation as naturally as an intelligent human – then it will be viewed as socially present. We will explore the issue of robot consciousness in greater detail later in this chapter.

Bonding Factor #2: Emotional Intelligence

Being perceived of as socially present is step one; however, that alone is not enough to foster widespread bonding between humans and robots. As emotional creatures, humans want to interact with other creatures that understand our emotions, react to them, and display emotions of their own. This means having emotional awareness, namely, the ability to interpret the emotions of others, and to respond to those emotions with realistic expressions of compassion and concern. In essence, the ability to augment one’s behavior based on the emotional state of others. This is often referred to as emotional intelligence, or as emotional quotient, EQ (Goleman 1995). We humans want to know that the other entity in the relationship understands us and feels for us. For humans, a fellow human completely devoid of emotions is a sociopath, and few people want to bond with sociopaths.

To assume the role of a true social companion, and to have humans bond with it, a robot must have the ability to do all three of the things mentioned earlier: (a) understand human emotions, (b) react to them appropriately, and (c) display emotions of its own. Human emotions are quite complex, and many humans struggle with emotional intelligence. However, robot creators are working hard to give robots this ability. Facial recognition software is advancing, allowing robots to not only recognize individuals (so they can greet them by name) but also recognize human emotions via the expressions we make. In this way, if a robot notices that its human companion is sad or stressed, it can react appropriately. Also, robots can increasingly understand subtleties in wording humans use, to try to interpret the feelings of their human companions. Emotions are often conveyed via sarcasm, and sarcasm is currently very difficult for artificial intelligence to interpret and understand, mainly because when people are sarcastic they mean the opposite of what they say (Such as: “It’s two degrees outside. How lovely!”). Artificial intelligence, currently, tends to take us literally. However, significant work is being done on helping artificial intelligence understand human sarcasm, by looking at context, tone of voice, facial expressions, past comments made by the same individual, and learning over time (Beckett 2018).

Among humans, the more experience we have with other individuals, the better we get at reading their emotions. Consider an elderly couple who have lived together for 50 years. They can often communicate to each other with the quickest of facial expressions. They have learned over the years to read each other’s feelings. Machine learning technology allows robots to do the same. The more they live with a human, the more they understand the subtle, or not so subtle, ways that an individual expresses his/her emotions. The ability of artificial intelligence to understand, react to, and display emotions is moving faster than many consumers might expect. According to Annette Zimmermann, Vice President of Research at Gartner, a global research and advising company active in the technology field, By 2022, your personal device will know more about your emotional state than your own family (Kleber 2018).

Both large, established companies, such as IBM, and smaller, newer companies, such as Emoshape (established in 2017) are investing heavily in this area. In the industry there are two terms commonly used to describe this ability: “emotional AI” and “affective computing.” Companies are investing heavily in this area because they understand emotional intelligence is essential if robots are to assume a larger and more meaningful role in the lives of humans (Goasduff 2018). Emoshape’s website (emoshape.com—as of November 20, 2019) talks about the company’s progress on emotion speech synthesis, emotional awareness, emotion reasoning, machine emotional intimacy…. The company is working hard on this technology because they realize, as they state on their website, that meaningful emotional interaction is core to removing the barrier to widespread adoption.

Pepper, from SoftBank Robotics (headquartered in Tokyo), is billed as a robot that can understand emotions. It does this via its facial recognition software that can interpret expressions, along with its speech recognition ability. It is far from perfect; however, Pepper does fairly well holding up its end of a fast-paced conversation with a human. Here is a video of Pepper in a conversation recorded in 2018 (www.youtube.com/watch?v=zJHyaD1psMc). Note Pepper’s excellent use of physical gestures, and the intonation it uses when speaking to avoid sounding monotonous. Pepper responds quickly and naturally to almost all the questions it is asked. However, during this conversation there are a few questions it cannot understand. Though Pepper is not perfect in its social role, this robot is the first step in a process that will certainly have significant improvements in the near future. It is a fairly impressive start.

Certainly, not every robot in our robot future needs to be emotionally advanced. There will likely be a large number of servant robots that do our menial work for us (vacuuming, mowing, etc.). We will likely view them as drones and have little interest in bonding with them. In fact, we may prefer to not bond with them because that might make us feel guilty having them work so hard for us. As Table 7.1 shows, most consumers (from one of our national surveys with a sample size of 493) do not believe that low level worker roles such as housework and yard work require any human emotions. However, for a host of other roles, consumers believe emotions are crucial to truly fulfilling that role. That includes serving as a companion or even acting as a greeter in a store or hospital (see Table 7.1)

Table 7.1 Roles people say require emotions

Bonding Factor #3: Individual Personality

Humans are social creatures who naturally prefer to live in communities. However, we are not ants or bees. We have no desire to live in mindless collectives with identical drones. Our intelligence has driven us to all be unique, and we want to interact with other unique individuals. Robots will need to present themselves as individuals. Humans will want to believe there is an “individual” with whom they are interacting. Much of this will be based on whether we humans perceive the robot as intelligent and autonomous (not, necessarily, conscious). If we go to our neighbor’s house and there is another robot that is exactly like ours—looks the same, talks the same, uses the exact same sentences, and so on—then human-robot bonding will be diminished. But if our robot expresses its own unique personality, then human-robot bonding becomes more likely.

Machine learning and artificial intelligence will allow robots to evolve uniqueness. According to Emoshape’s website, the company’s technology will lead to “…robots developing their own personalities, learning from human interactions.” (emoshape.com—as of November 20, 2019). Robot uniqueness will be expressed in ways similar to how humans express their uniqueness, including distinct perspectives on the world, distinct sense of humor, distinct personal interests, and distinct gestures and mannerisms.

Humans will project personality characteristics onto robots, driven largely by the robot’s physical characteristics and also the function it serves (Bennett 2018). Regarding physical characteristics, a cute humanoid appearance of a robot (as discussed in Chapter 6) will lead a human to project a more youthful and fun personality onto that robot. A cute robot will be seen as friendly and nonthreatening. A more mechanical and abstract appearance for a robot will lead a human to project a more cold and distant personality onto it. This type of robot will be viewed more along the lines of an inanimate instrument and less like an individual. Bonding will be more difficult with a heavily mechanical-looking robot versus a cute, humanoid robot. Size matters as well. A larger robot (equal or bigger than an adult human) may convey a more imposing and threatening personality than a robot that is only about half the size of an adult human.

The task robots perform will also impact our relationships with them. Robots performing higher level tasks, such as serving as a research librarian or surgeon, will be seen as having a more sophisticated personality. We may be more respectful and even deferential toward them. One the other hand, robots doing simple and repetitive tasks, like washing floors, will be viewed as having unassuming personalities and being of lower intelligence and hence of less interest to us.

Unique names and character backstories will help humans anthropomorphize robots (Darling, Nandy, and Breazeal 2015). Imagine, at some point in the future, spending a few days in a hospital recuperating from surgery. Rather than have a companion robot named “companion robot number three” spending time with you playing cards and conversing to help the days go by, you would interact with a robot name “Sammy.” Sammy’s name would be distinct from the other companion robots working at the hospital. It would portray itself as male, with a male voice. He might tell you he was created 5 years earlier, and has worked in two different cities. He may tell you what he liked and disliked about each city. He may tell you about his hobbies. These would be different from what the other companion robots at that institution might say about themselves. All of this would likely be fictional, but it would help develop a backstory making Sammy seem more like a unique individual, which would aid in your interactions with him.

Bonding Factor #4: Personal Benefit

For a bond to last, most humans need to feel they are getting something positive out of the relationship, such as a degree of personal satisfaction and fulfillment. That way, they will want to continue the relationship. Sometimes it is a feeling of being cared for. Other times, the relationship will make them laugh and find enjoyment in life. Or, the relationship is educational and helps the individual improve himself/herself. Given the advancing capabilities of robots, it seems likely that they can play these beneficial roles in the lives of humans in the near future.

However, is it ethical for robot manufacturers to create robots to play these social roles in people’s lives? Should robot manufacturers be intentionally making robots with which we humans will be prone to bond? On one hand, why not? If any technology can add positive elements in the lives of humans, why not support it? We can argue that much of modern technology has improved our quality of life, and social robots are just another step in this direction. Imagine an elderly person living alone who is unable to take care of a dog or cat. A companion robot might vastly improve her quality of life. However, on the other hand are the unintended consequences of modern technology, which can often be quite negative. For social robots and their bonding with humans, the biggest concern, as discussed by both our survey respondents and experts in Chapter 4, is that robot relationships can replace human relationships. And if human relationships are replaced in large enough numbers, the impact on society would be devastating. This is something robot manufacturers, companies that utilize robots, our political leaders and, in truth, all of us must be cognizant of as we march ahead into the robot future.

Robot Consciousness

We cannot explore the issue of human-robot bonding without considering the possibility of robot consciousness. First, we must define two terms commonly used in this area. First, sentience is defined as the ability of an entity to sense the world around it and respond to it (Armstrong 1981). Consciousness is a critical step beyond sentience. Consciousness refers to an entity not only being aware (sentient) of the environment around it but also being aware that it is aware—essentially self-awareness (Carruthers 2000). Robots today are equipped with the latest sensor technology; they can see, hear, smell, and feel. Hence, they have the ability to sense the environment around them (in many ways even better than humans), and their intelligent and autonomous abilities allow them to respond to it. However, all experts agree robots are not yet conscious. Even the most intelligent robots and most advanced AI programs have not yet achieved self-awareness.

But will the day come when robots do actually achieve self-awareness? This is a complex question, and there is no agreement among the experts. First, what exactly is consciousness? We have a definition of consciousness, but how do we operationalize and measure it? Second, are there levels of consciousness? Clearly, humans are self-aware. And there is general agreement that higher-level animals such as dogs are self-aware. But, how far down the animal chain can we go and still be sure the animals are self-aware? What about rats? Fish? Insects? There is debate regarding the degree of sentience and consciousness among lower-level animals and insects (Klein and Barron 2016; Woodruff 2017). A key challenge in this regard is referred to as the problem of other minds, which essentially refers to the fact that consciousness is a completely private and subjective experience, and hence we can never be sure of the level of consciousness of another entity—at least not with current technology. Given the ongoing discussion of levels of sentience and levels of consciousness as they apply to living creatures, it becomes quite complex to have this conversation regarding robots.

Currently, robotics and AI experts are divided regarding whether robots can ever truly become self-aware or not (Hildt 2019; Veliz 2016). Among those who believe robot self-awareness is eventually possible, there is no agreement if that can occur in the near future or if it is centuries away. If consciousness can occur in an artificial entity such as a robot, it will not be directly programmed in, at least not as we currently understand programming. If there is any hope at all for robot consciousness, it will likely come from the process of machine learning. Machine learning is the remarkable ability of highly advanced artificial intelligence programs to learn on their own. Machine learning allows robots to go beyond their initial programming. Based on their experiences interacting with the world around them, AI programs learn and advance on their own. For instance, an AI program recently developed its own successful strategies for the complicated Chinese game Go completely on its own (Gibney 2016). Go is considered far more complex than Chess, and an AI has recently beaten a human professional by learning, autonomously, how to improve its strategy over time. The AI program did this by repeatedly playing Go and teaching itself new and improved strategies that were never part of its original programming. If consciousness will ever develop in robots, it will likely come from something like this. If it is to happen, robots will “learn” self-awareness on their own. If robots become truly self-aware, humans will be even more likely to bond with them in even more meaningful ways because, in essence, robots will have become another race of conscious beings. However, this step is not necessary for robot-human bonding. As stated earlier, many humans will still develop social relationships and long-term bonding with robots even if robots do not develop true consciousness, as long as robots can deliver on the bonding factors reviewed above.

Three Levels of Human-Robot Relationships

Based on our own research and our review of existing scholarships and expert viewpoints, we propose that human interactions with robots in the coming decades will fall into one of three categories, based upon the robot’s capabilities, appearance, and roles. The three categories are:

1. Robots treated as tools
2. Robots treated as pets
3. Robots treated as humans

Category 1 Relationship: Robots Treated as Tools

Consumers will approach the more simplistic robots as tools; smart tools, but simply tools nonetheless. These will be the robots that are less intelligent and less interactive, playing entirely subservient, service roles. Humans will appreciate the work these robots do mowing the lawn, cleaning the carpets, and so on. However, we will devote minimal thought to them. There will be no meaningful relationship, no significant bonding (assuming these worker robots only work, and are not highly interactive social robots like the fictional maid Rosie on The Jetsons); and, on the positive side, no fear or apprehension. The interactions will be similar to how humans currently relate to their Roombas, room service delivery robots and their computers. This interaction will be the easiest for humans to accept, as we have already accepted smart tools in our lives.

Category 2 Relationship: Robots Treated as Pets

Robots that fall on the middle of the scale in terms of interactivity and intelligence will be treated as pets. They will have interactive capabilities such as speech, allowing us to have conversations with them. And they will be generally cute in appearance, fostering a degree of bonding. However, their limited intelligence will cause their human counterparts to see them as less than human. We will enjoy their company and we will interact with them. But we will always know (or at least believe) they are inferior to us in intellectual capability. Hence, the relationship will be similar to the one most humans have with their animal pets. These robots will be perceived of as generally nonthreatening because they will be seen as lesser entities compared to humans. Yet, these robots will be intelligent and interactive enough to become more interesting to humans. Current advanced companion robots discussed in Chapter 5 such as Sanbot Elf or Buddy currently, or with some further advancement will fall into this category. These relationships should form fairly easily for most humans, because we are still not treating the robots as equals; however, we are letting them get a bit closer to our hearts.

Category 3 Relationship: Robots Treated as Humans

Eventually, in the next couple of decades, robots will be so advanced in their interactive abilities that many humans will come to treat them as equals. This is because these robots will be able to communicate with us at the level of an intelligent human, and they will have emotional capabilities to react to our feelings and display sentiments of their own. They will gesture, joke, laugh, and cry just like our human friends and family members. They will achieve the four bonding factors noted earlier.

These robots will represent a substantial change to human society, something never before faced by our race—the introduction of a new species that can talk and interact with us in a manner that significantly mimics a fellow human. As stated earlier in the chapter, at first there will be awkwardness in the interactions between humans and advanced robots. Some humans will resist such interactions for quite a long time. But many people will, over time, embrace these relationships and increasingly treat these robots as equals, as “honorary” human beings. This means we will see them as individuals, with their own personalities. We will respect their opinions. We will say “please” and “thank you” to them. We will laugh with them, and enjoy their company. But, as with any human interactions, it will not all be positive. We will also sometimes fear them, envy them, and even hate them. We can foresee four general types of relationships within this “treat as human” category. Namely:

Category 3a: Treat as a Child. In this approach, humans will treat the robots as fellow humans, but more akin to how they would treat a child. This will be driven partly by a perception (rightly or wrongly) that the robot’s intelligence level is akin to that of a child. We may talk down to it. This may also be driven somewhat by the robot’s appearance. Current interactions with Pepper the Robot might fall somewhat into this area.

Category 3b: Treat as a Friend. This is the optimistic scenario. The robot is treated as a human, and the relationship is viewed by the human as highly positive.

Category 3c: Treat as a Competitor. This is a more pessimistic scenario. The robot is treated as a human; however, the relationship is viewed by the human as negative. It may be that the robot is a colleague at our workplace, but it is working more productively and hence getting more praise from our boss than we are. Or, the robot is a companion to a human friend, and getting more of that friend’s attention than we are.

Category 3d: Treat as a Master. This is the most pessimistic scenario—where the robots are viewed as humans, but seen as occupying a place of power and superiority vis-à-vis the individual. Perhaps the robot is an unforgiving boss. Perhaps the robot is running the local government.

Clearly, we are not there yet. But, highly advanced interactive robots are coming. The first wave is here. Pepper, by Softbank, is somewhat akin to a human child in its interactions. Sophia (introduced in Chapter 1) by Hanson Robotics is among the most advanced robots functioning today. She is getting closer and closer to a Category 3, Treat as Human, level. Whether her future iterations are perceived as friends, competitors, or masters is yet to be determined. Table 7.2 shows the three categories of human-robot interactions.

Table 7.2 Three categories of human-robot interactions

It is impossible to predict exactly how human-robot relationships will evolve, and over what timetable. One thing that is certain is that the robot future will be like nothing like we have ever seen before. Humanity has experienced massive technological changes over the past few generations. However, the robot future will be far different, and more profound, than anything we as a race have experienced so far.

Implications for Consumer Behavior and Marketing Strategy

Think about the robots used or sold by your company (when they eventually do), and what type of relationship you wish customers to form with them: as a tool, a pet, or as human. Think hard about what is needed to form each bond. To achieve a human-level bond, keep in mind the key conditions of social presence, emotional intelligence, individual personality, and mutual benefit.

Home companion robots should certainly aspire to reach a human-level bond. Yet, even a customer service robot in a store whose job it is to help customers find products needs to also meet this bar to a significant degree. Customers asking the store robot questions will need to think that they are speaking with an entity that is truly listening and actively communicating with them (social presence!). If customers get frustrated about anything, the robot should be able to understand and respond appropriately to those emotions (emotional intelligence!). Give your robots names and have them displayed on nametags. Have them each exhibit a slightly different personality (individuality!). While it might be acceptable for your business to have customer service robots that function mainly as “tools” (cold, matter-of-fact, computer-like machines), if your competitors have service robots that are better at forming human-level bonds, your company will be at a significant disadvantage. Many people enjoy a bit of friendly chitchat with store personnel, waiters, bartenders, nurses, receptionists, and on and on. If robots can effectively hold up their end of this interaction, it will be a more satisfactory experience for the customer versus interacting with cold, utilitarian machines.

This may be hard for some readers to absorb, since it is still early in the robot revolution. Bonding with robots feels far off. Well, it is coming sooner than many people think, and hopefully what you have read in this book has convinced you of this. As with any sociotechnological trend, the key is to stay ahead of it and understand it better than your competitors. While this issue may not be something you need to stay awake tonight worrying about, do not put it off forever. Further, even in here and now, the issue is relevant. Consider the Botlr robot (from Savioke) that is used as a room service delivery robot in many hotels (discussed in Chapter 4). After delivering the food, Botlr asks the hotel patron if everything is OK and once the customer responds in the affirmative, it sings a little happy tune and dances. These small touches help “humanize” a simple delivery robot and go a long way toward helping create a bond with the customer, and improving the customer experience. Think of small bonding touches now, and continue to brainstorm the bigger consumer-robot relationship issues that are certainly coming in the years and decades ahead.