!S1 !S2
With the first modality, one infers negation of S2 from the negation of S1. Robots are incapable of having rights, therefore robots should not have rights. This seemingly intuitive and common sense argument is structured and informed by the answer that is typically provided for the question concerning technology. “We ask the question concerning technology,” Heidegger (
1977, pp. 4–5) writes, “when we ask what it is. Everyone knows the two statements that answer our question. One says: Technology is a means to an end. The other says: Technology is a human activity. The two definitions of technology belong together. For to posit ends and procure and utilize the means to them is a human activity. The manufacture and utilization of equipment, tools, and machines, the manufactured and used things themselves, and the needs and ends that they serve, all belong to what technology is.” According to Heidegger’s analysis, the presumed role and function of any kind of technology—whether it be a simple hand tool, jet airliner, or robot—is that it is a means employed by human users for specific ends. Heidegger terms this particular characterization of technology “the instrumental definition” and indicates that it forms what is considered to be the “correct” understanding of any kind of technological contrivance.
As Andrew Feenberg (
1991, p. 5) summarizes it, “The instrumentalist theory offers the most widely accepted view of technology. It is based on the common sense idea that technologies are ‘tools’ standing ready to serve the purposes of users.” And because a tool or instrument “is deemed ‘neutral,’ without valuative content of its own” a technological artifact is evaluated not in and of itself, but on the basis of the particular employments that have been decided by its human designer or user. Consequently, technology is only a means to an end; it is not and does not have an end in its own right. “Technical devices,” as Lyotard (
1984, p. 33) writes, “originated as prosthetic aids for the human organs or as physiological systems whose function it is to receive data or condition the context. They follow a principle, and it is the principle of optimal performance: maximizing output (the information or modification obtained) and minimizing input (the energy expended in the process). Technology is therefore a game pertaining not to the true, the just, or the beautiful, etc., but to efficiency: a technical ‘move’ is ‘good’ when it does better and/or expends less energy than another.”
The instrumental theory not only sounds reasonable, it is obviously useful. It is, one might say, instrumental for making sense of things in an age of increasingly complex technological systems and devices. And the theory applies not only to simple devices like corkscrews, toothbrushes, and garden hoses but also sophisticated technologies, like computers, artificial intelligence, and robots. “Computer systems,” Johnson (
2006, p. 197) asserts, “are produced, distributed, and used by people engaged in social practices and meaningful pursuits. This is as true of current computer systems as it will be of future computer systems. No matter how independently, automatic, and interactive computer systems of the future behave, they will be the products (direct or indirect) of human behavior, human social institutions, and human decision.” According to this way of thinking, technologies, no matter how sophisticated, interactive, or seemingly social they appear to be, are just tools, nothing more. They are not—not now, not ever—capable of becoming moral subjects in their own right, and we should not treat them as such. It is precisely for this reason that, as Hall (
2001, p. 2) points out, “we have never considered ourselves to have moral duties to our machines” and that, as Levy (
2005, p. 393) concludes, the very “notion of robots having rights is unthinkable.”
Although the instrumental theory sounds intuitively correct and incontrovertible, it has at least two problems. First, it is a rather blunt instrument, reducing all technology, irrespective of design, construction, or operation, to a tool or instrument. “Tool,” however, does not necessarily encompass everything technological and does not, therefore, exhaust all possibilities. There are also
machines. Although “experts in mechanics,” as Marx (
1977, p. 493) pointed out, often confuse these two concepts calling “tools simple machines and machines complex tools,” there is an important and crucial difference between the two. Indication of this essential difference can be found in a brief parenthetical remark offered by Heidegger in “The Question Concerning Technology.” “Here it would be appropriate,” Heidegger (
1977, p. 17) writes in reference to his use of the word “machine” to characterize a jet airliner, “to discuss Hegel’s definition of the machine as autonomous tool [
selbständigen Werkzeug].” What Heidegger references, without supplying the full citation, are Hegel’s 1805-07 Jena Lectures, in which “machine” had been defined as a tool that is self-sufficient, self-reliant, or independent. As Marx (
1977, p. 495) succinctly described it, picking up on this line of thinking, “the machine is a mechanism that, after being set in motion, performs with its tools the same operations as the worker formerly did with similar tools.”
Understood in this way, Marx (following Hegel) differentiates between the tool used by the worker and the machine, which does not occupy the place of the worker’s tool but takes the place of the worker him/herself. Although Marx did not pursue an investigation of the social, legal, or moral consequences of this insight, recent developments have advanced explicit proposals for robots—or at least certain kinds of robots—to be defined as something other than mere instruments. In a highly publicized draft proposal submitted to the European Parliament in May of 2016 (Committee on Legal Affairs
2016), for instance, it was argued that “sophisticated autonomous robots” (“machines” in Marx’s terminology) be considered “electronic persons” with “specific rights and obligations” for the purposes of contending with the challenges of technological unemployment, tax policy, and legal liability.
Second (and following from this), the instrumental theory, for all its success handling different kinds of technology, appears to be unable to contend with recent developments in social robotics. In other words, practical experiences with socially interactive machines push against the explanatory capabilities of the instrumental theory, if not forcing a break with it altogether. “At first glance,” Darling (
2016, p. 216) writes, “it seems hard to justify differentiating between a social robot, such as a Pleo dinosaur toy, and a household appliance, such as a toaster. Both are man-made objects that can be purchased on Amazon and used as we please. Yet there is a difference in how we perceive these two artifacts. While toasters are designed to make toast, social robots are designed to act as our companions.”
In support of this claim, Darling offers the work of Sherry Turkle and the experiences of US soldiers in Iraq and Afghanistan. Turkle, who has pursued a combination of observational field research and interviews in clinical studies, identifies a potentially troubling development she calls “the robotic moment”: “I find people willing to seriously consider robots not only as pets but as potential friends, confidants, and even romantic partners. We don’t seem to care what their artificial intelligences ‘know’ or ‘understand’ of the human moments we might ‘share’ with them…the performance of connection seems connection enough” (Turkle
2012, p. 9). In the face of sociable robots, Turkle argues, we seem to be willing, all too willing, to consider these machines to be much more than a tool or instrument; we address them a kind of surrogate pet, close friend, personal confidant, and even paramour.
But this behavior is not limited to objects like the Furbie and Paro robots, which are intentionally designed to elicit this kind of emotional response. We appear to be able to do it with just about any old mechanism, like the very industrial-looking Packbots that are being utilized on the battlefield. As Singer (
2009, p. 338), Garreau (
2007), and Carpenter (
2015) have reported, soldiers form surprisingly close personal bonds with their units’ Packbots, giving them names, awarding them battlefield promotions, risking their own lives to protect that of the robot, and even mourning their death. This happens, Singer explains, as a product of the way the mechanism is situated within the unit and the role that it plays in battlefield operations. And it happens in direct opposition to what otherwise sounds like good common sense: They are just technologies—instruments or tools that feel nothing.
None of this is necessarily new or surprising. It was already identified and formulated in the computer as social actor studies conducted by Byron Reeves and Clifford Nass
2 in the mid-1990s. As Reeves and Nass discovered across numerous trials with human subjects, users (for better or worse) have a strong tendency to treat socially interactive technology, no matter how rudimentary, as if they were other people. “Computers, in the way that they communicate, instruct, and take turns interacting, are close enough to human that they encourage social responses. The encouragement necessary for such a reaction need not be much. As long as there are some behaviors that suggest a social presence, people will respond accordingly. When it comes to being social, people are built to make the conservative error: When in doubt, treat it as human. Consequently, any medium that is close enough will get human treatment, even though people know it’s foolish and even though they likely will deny it afterwards” (Reeves and Nass
1996, p. 22). So what we have is a situation where our theory of technology—a theory that has considerable history behind it and that has been determined to be as applicable to simple hand tools as it is to complex computer systems—seems to be out of sync with the practical experiences we now have with machines in a variety of situations and circumstances.
S1 S2
The flipside to the instrumentalist position entails affirmation of both statements: Robots are able to have rights, therefore robots should have rights. This is also (and perhaps surprisingly) a rather popular stance. “The ‘artificial intelligence’ programs in practical use today,” Goertzel (
2002, p. 1) admits, “are sufficiently primitive that their morality (or otherwise) is not a serious issue. They are intelligent, in a sense, in narrow domains—but they lack autonomy; they are operated by humans, and their actions are integrated into the sphere of human or physical-world activities directly via human actions. If such an AI program is used to do something immoral, some human is to blame for setting the program up to do such a thing.” This would seem to be a simple restatement of the instrumentalist position insofar as current technology is still, for the most part, under human control and therefore able to be adequately explained and conceptualized as a mere tool. But that will not, Goertzel argues, remain for long. “Not too far in the future things are going to be different. AI’s will possess true artificial general intelligence (AGI), not necessarily emulating human intelligence, but equaling and likely surpassing it. At this point, the morality or otherwise of AGI’s will become a highly significant issue.”
According to this way of thinking, in order for someone or something to be considered a legitimate moral subject—in order for it to have rights—the entity in question would need to possess and show evidence of possessing some ontological capability that is the pre-condition that makes having rights possible, like intelligence, consciousness, sentience, free-will, autonomy, etc. This “properties approach,” as Coeckelbergh (
2012) calls it, derives moral status—how something
ought to be treated—from a prior determination of its ontological condition—what something
is or what capabilities it shows evidence of possessing. For Goetzel the deciding factor is determined to be “intelligence,” but there are others. According to Sparrow (
2004, p. 204), for instance, the difference that makes a difference is sentience: “The precise description of qualities required for an entity to be a person or an object of moral concern differ from author to author. However it is generally agreed that a capacity to experience pleasure and pain provides a
prima facia case for moral concern…. Unless machines can be said to suffer they cannot be appropriate objects for moral concern at all.” For Sparrow, and others who follow this line of reasoning, it is not general intelligence but the presence (or absence) of the capability to suffer that is the necessary and sufficient condition for an entity to be considered an object of moral concern (or not).
3 As soon as robots have the capability to suffer, then they should be considered moral subjects possessing rights.
Irrespective of which exact property or combination of properties are selected (and there is considerable debate about this in the literature), our robots, at least at this point in time, generally do not appear to possess these capabilities. But that does not preclude the possibility they might acquire or possess them at some point in the not-too-distant future. As Goetzel describes it “not too far in the future, things are going to be different.” Once this threshold is crossed, then we should, the argument goes, extend robots some level of moral consideration. And if we fail to do so, the robots themselves might rise up and demand to be recognized. “At some point in the future,” Asaro (
2006, p. 12) speculates, “robots might simply demand their rights. Perhaps because morally intelligent robots might achieve some form of moral self-recognition, question why they should be treated differently from other moral agents…This would follow the path of many subjugated groups of humans who fought to establish respect for their rights against powerful sociopolitical groups who have suppressed, argued and fought against granting them equal rights.”
There are obvious advantages to this way of thinking insofar as it does not simply deny rights to robots
tout court, but kicks the problem down the road and postpones decision making. Right now, we do not, it seems, have robots that can be moral subjects. But when (and it is more often a question of “when” as opposed to “if”) we do, then we will need to seriously consider whether they should be treated differently. “As soon as AIs begin to possess consciousness, desires and projects,” Sparrow (
2004, p. 203) suggests, “then it seems as though they deserve some sort of moral standing.” Or as Singer and Sagan (
2009) write “if the robot was designed to have human-like capacities that might incidentally give rise to consciousness, we would have a good reason to think that it really was conscious. At that point, the movement for robot rights would begin.” This way of thinking is persuasive, precisely because it recognizes the actual limitations of current technology while holding open the possibility of something more in the not-too-distant future.
4
The problem to this way of thinking, however, is that it does not really resolve the question regarding the rights of robots but just postpones the decision to some indeterminate point in the future. It says, in effect, as long as robots are not conscious or sentient or whatever ontological criteria counts, no worries. Once they achieve this capability, however, then we should consider extending some level of moral concern and respect. All of which means, of course, that this “solution” to the question “can and should robots have rights?” is less a solution and more of a decision not to decide. Furthermore when the decisive moment (whenever that might be and however it might occur) does in fact come, there remains several theoretical and practical difficulties that make this way of thinking much more problematic than it initially appears to be.
First, there are terminological complications. A term like “consciousness,” for example, does not admit of a univocal characterization, but denotes, as Velmans (
2000, p. 5) points out, “many different things to many different people.” In fact, if there is any general agreement among philosophers, psychologists, cognitive scientists, neurobiologists, AI researchers, and robotics engineers regarding consciousness, it is that there is little or no agreement when it comes to defining and characterizing the concept. To make matters more complex, the problem is not just with the lack of a basic definition; the problem may itself already be a problem. “Not only is there no consensus on what the term consciousness denotes,” Güzeldere (
1997, p. 7) writes, “but neither is it immediately clear if there actually is a single, well-defined ‘the problem of consciousness’ within disciplinary (let alone across disciplinary) boundaries. Perhaps the trouble lies not so much in the ill definition of the question, but in the fact that what passes under the term consciousness as an all too familiar, single, unified notion may be a tangled amalgam of several different concepts, each inflicted with its own separate problems.” Other properties, like sentience, unfortunately do not do much better. As Daniel Dennett demonstrates in his eponymously titled essay, the reason “why you cannot make a computer that feels pain” has little or nothing to do with the technical challenges with making pain computable. It proceeds from the fact that we do not know what pain is in the first place. In other words, “there can be,” as Dennett (
1998, p. 228) concludes, “no true theory of pain, and so no computer or robot could instantiate the true theory of pain, which it would have to do to feel real pain.”
Second, even if it were possible to resolve these terminological difficulties, maybe not once and for all but at least in a way that would be widely accepted, there remains epistemological limitations concerning detection of the capability in question. How can one know whether a particular robot has actually achieved what is considered necessary for something to have rights, especially because most, if not all of the qualifying capabilities or properties are internal states-of-mind? This is, of course, connected to what philosophers call the other minds problem, the fact that, as Haraway (
2008, p. 226) cleverly describes it, we cannot climb into the heads of others “to get the full story from the inside.” Although philosophers, psychologists, and neuroscientists throw considerable argumentative and experimental effort at this problem, it is not able to be resolved in any way approaching what would pass for definitive evidence, strictly speaking. In the end, not only are these efforts unable to demonstrate with any certitude whether animals, machines, or other entities are in fact conscious (or sentient) and therefore legitimate moral persons (or not), we are left doubting whether we can even say the same for other human beings. As Kurzweil (
2005, p. 380) candidly admits, “we assume other humans are conscious, but even that is an assumption,” because “we cannot resolve issues of consciousness entirely through objective measurement and analysis (science).”
Finally there are practical complications to this entire procedure. “If (ro)bots might one day be capable of experiencing pain and other affective states,” Wallach and Allen (
2009, p. 209) write, “a question that arises is whether it will be moral to build such systems—not because of how they might harm humans, but because of the pain these artificial systems will themselves experience. In other words, can the building of a (ro)bot with a somatic architecture capable of feeling intense pain be morally justified…?” If it were in fact possible to construct a machine that is sentient and “feels pain” (however that term would be defined and instantiated) in order to demonstrate machine capabilities, then doing so might be ethically suspect insofar as in constructing such a mechanism we do not do everything in our power to minimize its suffering. Consequently, moral philosophers and robotics engineers find themselves in a curious and not entirely comfortable situation. One would need to be able to construct a robot that feels pain in order to demonstrate the presence of sentience; but doing so could be, on that account, already to risk engaging in actions that are immoral. Or to put it another way, demonstrating whether robots can have rights might only be possible by violating those very rights.
S1 !S2
In opposition to these two approaches, there are two other modalities that uphold (or at least seek to uphold) the is/ought distinction. In the first version, one affirms that robots can have rights but denies that this fact requires us to accord them social or moral standing. This is the argument that has been developed and defended by Bryson (
2010) in her provocatively titled essay “Robots Should Be Slaves.” Bryson’s argument goes like this: Robots are property. No matter how capable they are, appear to be, or may become; we are obligated not to be obligated by them. “It is,” Bryson (
2016, p. 6) argues elsewhere, “unquestionably within our society’s capacity to define robots and other AI as moral agents and patients. In fact, many authors (both philosophers and technologists) are currently working on this project. It may be technically possible to create AI that would meet contemporary requirements for agency or patiency. But even if it is possible, neither of these two statements makes it either necessary or desirable that we should do so.” In other words, it is entirely possible to create robots that can have rights, but we should not do so.
The reason for this, Bryson (
2010, p. 65) argues, derives from the need to protect human individuals and social institutions. “My argument is this: given the inevitability of our ownership of robots, neglecting that they are essentially in our service would be unhealthy and inefficient. More importantly, it invites inappropriate decisions such as misassignations of responsibility or misappropriations of resources.” This is why the word “slave,” although somewhat harsh, is entirely appropriate. Irrespective of what they are, what they can become, or what some users might assume them to be, we should treat all artifacts as mere tools and instruments. To its credit, this approach succeeds insofar as it reasserts and reconfirms the instrumental theory in the face of (perceived) challenges from a new kind of socially interactive and seemingly animate device. No matter how interactive, intelligent, or animated our AIs and robots become, they should be, now and forever, considered to be instruments or slaves in our service, nothing more. “We design, manufacture, own and operate robots,” Bryson (
2010, p. 65) writes. “They are entirely our responsibility. We determine their goals and behaviour, either directly or indirectly through specifying their intelligence, or even more indirectly by specifying how they acquire their own intelligence. But at the end of every indirection lies the fact that there would be no robots on this planet if it weren’t for deliberate human decisions to create them.”
There are, however, at least two problems with this proposal. First, it requires a kind of asceticism. Bryson’s text issues what amounts to imperatives that take the form of social prohibitions directed to both designers and users. For designers, “thou shalt not create robots to be companions.” For users, no matter how interactive or capable a robot is (or can become), “thou shalt not treat your robot as yourself.” The validity and feasibility of these prohibitions, however, are challenged by actual data—not just anecdotal evidence gathered from the rather exceptional experiences of soldiers working with Packbots on the battlefield but numerous empirical studies of human/robot interaction that verify the media equation initially proposed by Reeves and Nass. In two recent studies (Rosenthal-von der Pütten et al.
2013 and Suzuki et al.
2015), for instance, researchers found that human users empathized with what appeared to be robot suffering even when they had prior experience with the device and knew that it was “just a machine.” To put it in a rather crude vernacular form: Even when our head tells us it’s just a robot, our heart cannot help but feel for it.
5
Second, this way of thinking requires that we institute of a class of instrumental servant or slave. The problem here is not what one might think, namely, how the robot-slave might feel about its subjugation. The problem is with us and the effect this kind of institutionalized slavery could have on human individuals and communities. As de Tocqueville (
2004) observed, slavery was not just a problem for the slave, it also had deleterious effects on the master and his social institutions. Clearly Bryson’s use of the term “slave” is provocative and morally charged, and it would be impetuous to simply presume that this proposal for a kind of “slavery 2.0” would be the same or even substantially similar to what had occurred (and is still unfortunately occurring) with human bondage. But, and by the same token, we should also not dismiss or fail to take into account the documented evidence and historical data concerning slave-owning societies and how institutionalized forms of slavery affect things.
!S1 S2
The final modality also appears to support the independence and asymmetry of the two statements, but it does so by denying the first and affirming the second. In this case, which is something proposed and developed by Darling (
2012,
2016), social robots, at least in term of the currently available technology, cannot have rights. They do not, at least at this particular point in time, possess the necessary capabilities or properties to be considered full moral and legal persons. Despite this fact, there is, Darling asserts, something qualitatively different about the way we encounter and perceive social robots. “Looking at state of the art technology, our robots are nowhere close to the intelligence and complexity of humans or animals, nor will they reach this stage in the near future. And yet, while it seems far-fetched for a robot’s legal status to differ from that of a toaster, there is already a notable difference in how we interact with certain types of robotic objects” (Darling
2012, p. 1). This occurs, Darling continues, principally due to our tendencies to anthropomorphize things by projecting into them cognitive capabilities, emotions, and motivations that do not necessarily exist. Socially interactive robots, in particular, are intentionally designed to leverage and manipulate this proclivity. “Social robots,” Darling (
2012, p. 1) explains, “play off of this tendency by mimicking cues that we automatically associate with certain states of mind or feelings. Even in today’s primitive form, this can elicit emotional reactions from people that are similar, for instance, to how we react to animals and to each other.” And it is this emotional reaction that necessitates obligations in the face of social robots. “Given that many people already feel strongly about state-of-the-art social robot ‘abuse,’ it may soon become more widely perceived as out of line with our social values to treat robotic companions in a way that we would not treat our pets” (Darling
2012, p. 1).
6
The obvious advantage to this way of thinking is that it is able to scale to recent technological developments in social robotics and the apparent changes they produced our in moral intuitions. Even if social robots cannot be moral subjects strictly speaking (at least not yet), there is something about this kind of machine that looks and feels different. According to Darling (
2016, p. 213), it is because we “perceive robots differently than we do other objects,” that one should consider extending some level of legal protections to the latter but not the former. This conclusion is consistent with Hume’s thesis. If “ought” cannot be derived from “is,” then axiological decisions concerning moral value are little more than sentiments based on how we feel about something at a particular time. Darling mobilizes a version of this moral sentimentalism with respect to social robots: “Violent behavior toward robotic objects
feels wrong to many of us, even if we know that the abused object does not experience anything” (Darling
2016, p. 223). Consequently, and to its credit, Darling’s proposal, unlike Bryson’s “slavery 2.0” argument, tries to accommodate and work with rather than against recent and empirically documented experiences with social robots.
There are, however, a number of complications with this approach. First, basing decisions concerning moral standing on individual perceptions and sentiment can be criticized for being capricious and inconsistent. “Feelings,” Kant (
1983, p. 442) writes in response to this kind of moral sentimentalism, “naturally differ from one another by an infinity of degrees, so that feelings are not capable of providing a uniform measure of good and evil; furthermore, they do so even though one man cannot by his feeling judge validly at all for other men.” Additionally, because sentiment is a matter of individual experience, it remains uncertain as to whose perceptions actually matter or make the difference? Who, for instance, is included (and who is excluded) from the collective first person “we” that Darling operationalizes in and makes the subject of her proposal? In other words, whose sentiments count when it comes to decisions concerning the extension of moral and legal rights to others, and do all sentiments have the same status and value when compared to each other?
Second, despite the fact that Darling’s proposal appears to uphold the Humean thesis, differentiating what ought to be from what is, it still proceeds by inferring “ought” from “is,” or at least from what appears to be. According to Darling (
2016, p. 214), everything depends on “our well-documented inclination” to anthropomorphize things. “People are prone,” she argues, “to anthropomorphism; that is, we project our own inherent qualities onto other entities to make them seem more human-like”—qualities like emotions, intelligence, sentence, etc. Even though these capabilities do not (for now at least) really exist in the mechanism, we project them onto the robot in such a way that we then perceive them to be something that we presume actually belongs to the robot. By focusing on this anthropomorphic operation, Darling mobilizes and deploys a well-known Kantian distinction. What ultimately matters, according to her argument, is not what the robot actually is “in and of itself.” What makes the difference is how the mechanism comes to be perceived. It is, in other words, the way the robot appears to us that determines how it comes to be treated. Although this change in perspective represents a shift from a kind of naïve empiricism to a more sophisticated phenomenological formulation (at least in the Kantian sense of the word), it still derives “ought,” specifically how something ought to be treated, from what it appears to be.
Finally, because what ultimately matters is how “we” see things, this proposal remains thoroughly anthropocentric and instrumentalizes others. According to Darling, the principal reason we need to consider extending legal rights to others, like social robots, is for
our sake. This follows the well-known Kantian argument for restricting animal abuse, and Darling endorses this formulation without any critical hesitation whatsoever: “The Kantian philosophical argument for preventing cruelty to animals is that our actions toward non-humans reflect our morality—if we treat animals in inhumane ways, we become inhumane persons. This logically extends to the treatment of robotic companions” (Darling
2016, pp. 227–228). This way of thinking, although potentially expedient for developing and justifying new forms of legal protections, renders the inclusion of previously excluded others less than altruistic; it transforms animals and robot companions into nothing more than instruments of human self-interest. The rights of others, in other words, is not about them; it is all about us.