1 Thinking the Unthinkable

1.3 Robot Rights or the Unthinkable

“The notion of robot rights,” as Seo-Young Chu (2010, 215) points out, “is as old as is the word ‘robot’ itself. Etymologically the word ‘robot’ comes from the Czech word ‘robota,’ which means ‘forced labor.’ In Karel Čapek’s 1921 play R.U.R., which is widely credited with introducing the term ‘robot,’ a ‘Humanity League’ decries the exploitation of robot slaves—‘they are to be dealt with like human beings,’ one reformer declares—and the robots themselves eventually stage a massive revolt against their human makers.” For many researchers and developers, however, the very idea of “robot rights” is simply unthinkable. “At present,” Sohail Inayatullah and Phil McNally (1988, 123) once wrote, “the notion of robots with rights is unthinkable whether one argues from an ‘everything is alive’ Eastern perspective or ‘only man is alive’ Western perspective.” But why? How and why is the mere idea of “robot rights” inconceivable? What does it mean, when the mere thought of something is already decided and declared to be unthinkable?

1.3.1 Ridiculous Distractions

Considering the impact of his earlier (1988) essay with Phil McNally, Inayatullah (2001, 1) provided the following reflection concerning the question of robot rights:

Many years ago in the folly of youth, I wrote an article with a colleague titled, “The Rights of Robots.” It has been the piece of mine most ridiculed. Pakistani colleagues have mocked me saying that Inayatullah is worried about robot rights while we have neither human rights, economic rights or rights to our own language and local culture—we have only the “right to be brutalized by our leaders” and Western powers. Others have refused to enter in collegial discussions on the future with me as they have been concerned that I will once again bring up the trivial. As noted thinker Hazel Henderson said, I am happy to join this group—an internet listserv—as long as the rights of robots are not discussed. But why the ridicule and anger?

As Inayatullah reports from his own experience, the mere idea of robot rights—not even advocacy for rights but the suggestion that the question might require some careful consideration—has produced anger, mockery, and ridicule in colleagues and noted thinkers. In the face of this seemingly irrational response, Inayatullah asks a simple but important question, “why the ridicule and anger?” In response to this, he calls upon the work of Christopher Stone, who once pointed out that any extension of rights to previously excluded populations—like animals or the environment—is always, at least initially, considered unthinkable and the target of ridicule. “Throughout legal history,” Stone (1974, 6–7) writes, “each successive extension of rights to some new entity has been, theretofore, a bit unthinkable. … The fact is, that each time there is a movement to confer rights onto some new ‘entity,’ the proposal is bound to sound odd or frightening or laughable.” This mode of response is, for example, evident in the editors’ note appended to the beginning of a reprint of Marvin Minsky’s “Alienable Rights”: “Recently we heard some rumblings in normally sober academic circles about robot rights. We managed to keep a straight face as we asked Marvin Minsky, MIT’s grand old man of artificial intelligence, to address the heady question” (Minsky 2006, 137).³ From the outset, the question of robot rights is presumed to be ludicrous, so much so that one struggles to keep a straight face in the face of the mere idea. Whatever the exact reason for this kind of reaction—whether it be fear, resentment, prejudice, etc.⁴—the effect of ridicule is to discredit or to dismiss the very idea.

Similar responses have been registered and documented elsewhere in the literature. In 2006, the UK Office of Science and Innovation’s Horizon Scanning Centre commissioned and published a report consisting of, as Raya Jones (2016, 36) explains, “246 summary papers that predict emerging trends in science, health, and technology.” The papers, or “scans” as Sir David King referred to them, were “aimed at stimulating debate and critical discussion to enhance government’s short and long term policy and strategy” (BBC News 2006). Despite the fact that these scans (which had been contracted to and developed by Ipsos MORI in 2006) addressed a number of potentially controversial subjects, only one was “singled out to make headlines” (Jones 2016, 36). Ipsos MORI titled the article: “Robo-rights: Utopian Dream or Rise of the Machines?” (which was called “Robot Rights: Utopian Dream or Rise of the Machines?” in Winfield 2007 and “Robo-rights: Utopian Dream or Rise of the Machine?” in Bensoussan and Bensoussan 2015). What caught everyone’s attention was the following statement: “If artificial intelligence is achieved and widely deployed (or if they can reproduce and improve themselves) calls may be made for human rights to be extended to robots” (Marks 2007; Ipsos MORI 2006). The statement was short and the report itself, which was “a little over 700 words” (Winfield 2007), did not supply much in terms of actual details or evidence. But it definitely had impact. It was covered by and widely reported in the popular press, with attention grabbing headlines like “UK report says robots will have rights” (Davoudi 2006) and “Robots could demand legal rights” (BBC News 2006).

The response from researchers and experts in the field was highly critical.⁵ As Robert M. Geraci (2010, 190) explained: “The Ipsos MORI document was reviled by some scientists in the UK. Rightly criticizing the study for its minimal research documentation, Owen Holand [SIC], Alan Winfield, and Noel Sharkey claim the Ipsos MORI document directs attention away from the real issue, especially military robots and responsibility for autonomous robot-incurred death or damage (Henderson 2007). The scientists gathered at the Dana Centre in London in April of 2007 to share their ideas with the public.” This account of the sequence of events is not entirely accurate. Holland, Sharkey, and Winfield—along with Kathleen Richardson, whose name is not, for some reason, included in the published news stories—shared their comments about the Ipsos MORI report one day before the Dana Centre event during a press briefing at the Science Media Centre (SMC). In their remarks to the press, Holland, Sharkey, and Winfield—who together with Frank Burnett were co-PIs on the “Walking With Robots” research program, a multi-year public information project funded by the UK’s Engineering and Physical Sciences Research Council (2005)—not only criticized the report but reportedly dismissed the concept of robot rights as a “waste of time and resources” (Geraci 2010, 190) and a “red herring that has diverted attention from more pressing ethical issues” (Henderson 2007). As Holland explained, the report was “very shallow, superficial, and poorly informed. I know of no one within the serious robotics community who would use that phrase, ‘robot rights.’” (Henderson 2007). And according to Sharkey “the idea of machine consciousness and rights is a distraction, it’s fairy tale stuff. We need proper informed debate, about the public safety about for instance the millions of domestic robots that are predicted to be arriving in the next few years” (Henderson 2007 and Marks 2007).

By all accounts, the report was insufficient, perhaps necessarily so, since it was not a peer-reviewed scientific study but a short thought-piece designed to motivate public discussion. There is, however, a bit of oversteer in these critical responses. In addition to directing criticism at the document and its perceived deficiencies, the respondents also characterized the entire subject of robot rights as a fantastic distraction from the serious work that needs to be done and therefore something no serious researcher would think about, let alone say out loud. In all fairness, it is entirely possible that the remarks were motivated by a well-intended effort to try to direct public attention to the “real problems” and away from dramatic media-hype and futurist speculation. And this is undoubtedly one of the important and necessary objectives of science education. But even if we grant this, the comments demonstrate something of a dismissive response to the question of robot rights and its relative importance (or lack thereof) in current research programs. Even when taking a more measured and less polemic approach, as Winfield did in a blog post from February of 2007, the “rights question” is, if not dismissed, then indefinitely delayed by being put aside or pushed off into an indeterminate future.

Ok, let’s get real. Do I think robots will have (human) rights within 20–50 years? No, I do not. Or to put it another way, I think the likelihood is so small as to be negligible. Why? Because the technical challenges of moving from insect-level robot intelligence, which is more or less where we are now, to human-level intelligence are so great. Do I think robots will ever have rights? Well, perhaps. In principle I don’t see why not. Imagine sentient robots, able to fully engage in discourse with humans, on art, philosophy, mathematics; robots able to empathi[z]e or express opinions; robots with hopes, or dreams. Think of Data from Star Trek. It is possible to imagine robots smart, eloquent and persuasive enough to be able to argue their case but, even so, there is absolutely no reason to suppose that robot emancipation would be rapid, or straightforward. After all, even though the rights of man as now generally understood were established over 200 years ago, human rights are still by no means universally respected or upheld. Why should it be any easier for robots? (Winfield 2007).

In this post, Winfield does not dismiss the concept of robot rights per se, but comes at it directly. In the short term (20–50 years), it appears to be, if not impossible, then at least highly improbable. In the long run, it may be something that is, in principle at least, not impossible. This possibility, however, is projected into a distant hypothetical future on the basis of the more exclusive Will Theory of rights, i.e., that robots would be “smart, eloquent and persuasive enough to be able to argue their case.” The question concerning robot rights, therefore, is not “unthinkable”; it is just unlikely and practically not worth the effort at this particular point in time. As Holland had succinctly stated it during the press briefing, “it’s really premature I think to discuss robot rights” (Randerson 2007).

A similar decision is operationalized in the official response issued by the Foundation for Responsible Robotics to the European Parliament’s recently published recommendations to the European Commission concerning robots and AI (Committee on Legal Affairs 2016): “Some of the wording in the document is disappointingly based on Science Fiction. Notions of robot rights and robot citizenship are really a distraction from the main thrust of the document that only served to catch the wrong type of media attention” (Sharkey and Fosch-Villaronga 2017). Likewise, Luciano Floridi (2017, 4) suggests that thinking and talking about the “counterintuitive attribution of rights” to robots is a distraction from the serious work of philosophy. “It may be fun,” Floridi (2017, 4) writes, “to speculate about such questions, but it is also distracting and irresponsible, given the pressing issues we have at hand. The point is not to decide whether robots will qualify someday as a kind of persons, but to reali[z]e that we are stuck within the wrong conceptual framework. The digital is forcing us to rethink new solutions for new forms of agency. While doing so we must keep in mind that the debate is not about robots but about us, who will have to live with them, and about the kind of infosphere and societies we want to create. We need less science fiction and more philosophy.” According to Floridi, it might be fun to speculate about things like robot rights, but this kind of playful thinking is a distraction from the real work that needs to be done and quite possibly is an irresponsible waste of time and effort that should be spent on more serious philosophical endeavors.

1.3.2 Justifiable Exclusions

In other situations the question of robot rights is not immediately ridiculed or dismissed; it is given some brief attention only to be bracketed or carefully excluded as an area that shall not be given further thought. One of the best illustrations of this is deployed by Michael Anderson, Susan Leigh Anderson, and Chris Armen (2004) in the agenda-setting paper that launched the new field of machine ethics. For Anderson et al., machine ethics (ME) is principally concerned with the consequences of the decisions and behaviors of machines toward human beings. For this reason, Anderson et al. immediately differentiate ME from two related efforts: “Past research concerning the relationship between technology and ethics has largely focused on responsible and irresponsible use of technology by human beings, with a few people being interested in how human beings ought to treat machines.” ME is introduced and defined by being differentiated from two other ways of considering the relationship between technology and ethics. The first is computer ethics, which is concerned, as Anderson and company correctly point out, with questions of human action through the instrumentality of computers and related information systems. In clear distinction from these efforts, machine ethics seeks to enlarge the scope of moral agents by considering the ethical status and actions of machines. As Anderson and Anderson (2007, 15) describe it in a subsequent publication, “the ultimate goal of machine ethics, we believe, is to create a machine that itself follows an ideal ethical principle or set of principles.” The other exclusion involves situations regarding rights, or “how human beings ought to treat machines.” This also does not fall under the purview of ME, and Anderson, Anderson, and Armen explicitly mark it as something to be set aside by their own endeavors. Although the “question of whether intelligent machines should have moral standing,” Susan Leigh Anderson (2008, 480) writes in another article, appears to “loom on the horizon,” ME deliberately and explicitly pushes this issue to the margins, leaving it for others to consider. Here we can perceive how that which is marginalized appears; it makes an appearance in the text only insofar as it is set aside and not given further consideration. It appears, as Derrida (1982, 65) describes it, by way of the trace of its erasure. This setting-aside or exclusion leaves a residue in the text—a mark within the text of something being excluded—that remains legible and can be read.

A similar decision is deployed in Wendell Wallach and Colin Allen’s Moral Machines (2009), and is clearly evident in their choice of the term “artificial moral agent,” or AMA, as the protagonist of the analysis. This term immediately focuses attention on the question of responsibility. Despite this exclusive concern, however, Wallach and Allen (2009, 204–207) do eventually give brief consideration to the issue of rights. Extending the concept of legal responsibility to AMAs is, in Wallach and Allen’s opinion, something of a no-brainer: “The question whether there are barriers to designating intelligent systems legally accountable for their actions has captured the attention of a small but growing community of scholars. They generally concur that the law, as it exists, can accommodate the advent of intelligent (ro)bots. A vast body of law already exists for attributing legal personhood to nonhuman entities (corporations). No radical changes in the law would be required to extend the status of legal person to machines with higher-order faculties, presuming that the (ro)bots were recognized as responsible agents” (Wallach and Allen 2009, 204). According to Wallach and Allen’s estimations, a decision concerning the legal status of AMA’s should not pose any significant problems. Most scholars, they argue, already recognize that this is adequately anticipated by and already has a suitable precedent in available legal and judicial practices, especially as it relates to the corporation.

What is a problem, in their eyes, is the flip side of legal responsibility—the question of rights. “From a legal standpoint,” Wallach and Allen continue, “the more difficult question concerns the rights that might be conferred on an intelligent system. When or if future artificial moral agents should acquire legal status of any kind, the question of their legal rights will also arise” (Wallach and Allen 2009, 204). Although noting the possibility and importance of the question, at least as it would be characterized in legal terms, they do not pursue its consequences very far. In fact, they mention it only to defer it to another kind of question—a kind of investigative bait and switch: “Whether or not the legal ins and outs of personhood can be sorted out, more immediate and practical for engineers and regulators is the need to evaluate AMA performance” (Wallach and Allen 2009, 206). Consequently, Wallach and Allen conclude Moral Machines by briefly gesturing in the direction of a consideration of robot rights only to refer this question back to the issue of agency and performance measurement. In this way, then, they briefly address the rights question only to immediately recoil from the complications it entails, namely, the persistent philosophical problem of having to sort out “the legal ins and outs of personhood.” Although not simply passing over the question of robot rights in silence, Wallach and Allen, like Anderson et al., only mention it in order to postpone or otherwise exclude the issue from further consideration.

In other cases, the exclusion is not even explicitly marked or justified; it is simply declared and enacted. The EURON RoboEthics Roadmap (Veruggio 2006), for instance, does not have anything to say (not even a negative exclusion) about the social standing or rights of robots. As Yueh-Hsuan Weng, Chien-Hsun Chen, and Chuen-Tsai Sun (2009, 270) report, “the Roboethics Roadmap does not consider potential problems associated with robot consciousness, free will, and emotions” but takes an entirely human-centric approach. The Roadmap, then, appears to pursue what is a rather reasonable path. The scope of the report is limited to a decade, and, for this reason, the authors of the document set aside the more speculative questions concerning future achievements with robot capabilities. Though robot rights is thinkable in principal, in practice it is left off the table as a matter of setting some practical limits to the investigation.

In terms of scope, we have taken into consideration—from the point of view of the ethical issues connected to Robotics—a temporal range of a decade, in whose frame we could reasonably locate and infer—on the basis of the current State-of-the-Art in Robotics—certain foreseeable developments in the field. For this reason, we consider premature—and have only hinted at—problems inherent in the possible emergence of human functions in the robot: like consciousness, free will, self-consciousness, sense of dignity, emotions, and so on. Consequently, this is why we have not examined problems—debated in the literature—like the need to consider robot as our slaves, or the need to guarantee them the same respect, rights and dignity we owe to human workers. (Veruggio 2006, 7)

In this document, the question of rights is deliberately set aside as a matter that is premature insofar as it requires advancements in robotics (i.e., consciousness, free will, emotions, etc.) that remain futuristic, debatable, and outside the scope of the study’s ten-year window of opportunity.

Not surprisingly, a similar sounding decision is instituted by Gianmarco Veruggio (the project coordinator of the EURON Roboethics Roadmap) and Fiorella Operto, in a subsequent contribution to the Springer Handbook of Robotics:

In terms of scope, we have taken into consideration—from the point of view of the ethical issues connected to robotics—a temporal range of two decades, in whose frame we could reasonably locate and infer—on the basis of the current state-of-the-art in robotics—certain foreseeable developments in the field. For this reason, we consider premature—and have only hinted at—problems related to the possible emergence of human qualities in robots: consciousness, free will, self-consciousness, sense of dignity, emotions, and so on. Consequently, this is why we have not examined problems—debated in some other papers and essays—like the proposal to not behave with robots like with slaves, or the need to guarantee them the same respect, rights, and dignity we owe to human workers. Likewise, and for the same reasons, the target of roboethics is not the robot and its artificial ethics, but the human ethics of the robots’ designers, manufacturers, and users. Although informed about the issues presented in some papers on the need and possibility to attribute moral values to robots’ decisions, and about the chance that in the future robots might be moral entities like—if not more so than—human beings, the authors have chosen to examine the ethical issues of the human beings involved in the design, manufacturing, and use of the robots. (Veruggio and Operto 2008, 1501)

Here again, Veruggio (who is credited with initially introducing and defining the project of roboethics) and Operto limit consideration to the near-term—two decades. They therefore explicitly set aside and postpone any consideration of robots as either independent agents or patients and focus their attention on a kind of elaboration of computer ethics, that is, the ethical issues encountered by human beings in the design, manufacture, and use of robots.

A similar future-oriented disclaimer is deployed in Michael Nagenborg et al.’s critical investigation of robot regulation across Europe. In this text, the authors consider a number of issues regarding “responsibility and autonomous robots,” but they set aside the question of rights as not pertinent to the matter at hand: “In the context of this paper we therefore assume that artificial entities are not persons and not the bearers of individual, much less civil, rights. This does not imply that in a legal and ethical respect we could not grant a special status to robots nor that the development of artificial persons can be ruled out in principle. However, for the present moment and the nearer future we do not see the necessity to demand a fundamental change of our conception of legality. Thus, we choose a human-centred approach” (Nagenborg et al. 2008, 350). Nagenborg and company can safely and justifiably make this assumption, which to their credit they recognize and identify as an “assumption,” because robots are widely regarded as things that are not “bearers of individual, much less civil, rights” (2008, 350). This could, the authors recognize, change at some point in the future, and that change, in distinction to what Winfield (2007) had argued, could be made and justified in term of an Interest Theory insofar as we would decide to “grant … special status to robots.” But that future is so distant that it is not even worth serious consideration at this particular moment.

In other situations, exclusion occurs not as a deliberate postponement but as a product of the means of inclusion. On the legal side, for instance, the matter of “robot rights” is already effectively disabled by the way we have decided to incorporate robots into the law and how that mode of incorporation differs from another, similarly situated entity. As Gunther Teubner (2006, 521) explains:

Law is opening itself for the entry of new juridical actors—animals and electronic agents. The difference in outcomes, however, are striking. … Animal rights and similar constructs create basically defensive institutions. Paradoxically, they incorporate animals in human society in order to create defenses against destructive tendencies of human society against animals. The old formula of social domination of nature is replaced by the new social contract with nature. For electronic agents, the exact opposite is true. Their legal personification, especially in economic and technological contexts, creates aggressive new action centers as basic productive institutions. Here, their inclusion into society does not protect the new actors—just the opposite. It is society that needs to defend itself against the new actors.

Both robots and animals can be considered “the excluded other” of human social institutions, especially legal institutions. The way these two previously marginalized entities come to be included in law, however, is remarkably different. Animals—at least some species of animals—now have rights that are legally recognized in an effort to defend the welfare and interests of these vulnerable entities from “the destructive tendencies of human society.” Robots, or what Teubner calls “electronic agents” (which is already a mark of difference, insofar as animals are not describe as “agents” but occupy the place of “patient”), come to be included in legal matters for the opposite reason. Robots figure in law not for the sake of protecting the rights of the machine from human exploitation and misuse but for the sake of defending human society from potentially disastrous robot decision-making and action. Animals have rights; robots have responsibilities. It is, therefore, not the case that robots are simply excluded from the law, but that the means of their inclusion is accomplished in such a way that already renders the question of rights inoperative and effectively unthinkable.

1.3.3 Literal Marginalization

In still other situations, the question of robot rights is not excluded per se but included by being pushed to the margins of proper consideration. What is marginal is noted and given some attention, but it is deliberately marked as something located on the periphery of what is determined to be really important and worthy of investigation. Indication of the marginal position or at least the peripheral status of the “rights question” can be found in Wendell Wallach and Peter Asaro’s collection of foundational texts in robot and machine ethics (2017). Out of the thirty-five essays that are assembled in this volume, which is intended “to provide a roadmap of core concerns” by bring together a “careful selection of articles that have or will play an important role in ongoing debates” (Wallach and Asaro 2017, xii), only one essay, the final one, takes up and engages the question of rights. The position of this text within the sequence of the chapters, along with the introductory explanation provided by the editors, situate the inquiry into robot rights as a kind of future-oriented epilogue:

Perhaps the most futuristic and contested question in robot ethics is whether robots may someday have the capacity or capability to be the bearer of rights, and in virtue of what criteria will they deserve those rights? Gunkel (2014) defends the view that there may be robots deserving of the legal and moral status of having rights, and challenges the current standards of evaluating which entities (corporations, animals, etc.) deserve moral standing. Such futuristic legal questions underscore the way in which artificial agents presently serve as a thought experiment for legal theorists and a vehicle for reflections on the nature of legal agency and rights. (Wallach and Asaro 2017, 12)

In pointing this out, I am not complaining. It was both an honor and privilege to have this text included in the project. But the manner of its inclusion—the fact that it is the sole essay addressing the question of robot rights and that it comes “at the end of the story”—indicates the extent to which the “rights question” has been and remains a kind of marginalized afterthought in the literature of robot and machine ethics.

This literal maginalization is also evident in Patrick Lin et al.’s Robot Ethics, which has now appeared in two versions. In the first iteration, Robot Ethics (Lin, Abney, and Bekey 2012), the question concerning robot rights is explicitly addressed in the book’s final section, “VII Rights and Ethics.” This section consists of three essays from Rob Sparrow (2012), Kevin Warwick (2012), and Anthony F. Beavers (2012), and it is introduced with the following editorial note:

The preceding chapter 18 examined the ethics of robot servitude: Is it morally permissible to enforce servitude on robots, sometimes termed “robot slavery”? But to call such servitude “slavery” is inapt, if not seriously misleading, if robots have no will of their own—if they lack the sort of freedom we associate with moral personhood and moral rights. However, could robots someday gain what it takes to become a rights holder? What exactly is it that makes humans (but not other creatures) here on Earth eligible for rights? Is there a foreseeable future in which robots will demand their own “Emancipation Proclamation”? (Lin, Abney, and Bekey 2012, 299)

Like Wallach and Asaro, Lin et al. conclude Robot Rights (version 1, 2012) by giving consideration to a set of questions involving somewhat distant matters that could arise in the foreseeable future. “Robot rights,” although not a central concern of Robot Ethics at the present time, is added on at the end as a kind of indication of “future directions” for possible research: “Thus, after studying issues related to programming ethics and specific areas of robotic applications, in part VII our focus zooms back out to broader, more distant concerns that may arise with future robots” (Lin, Abney, and Bekey 2012, 300). But in the second version of the book, Robot Ethics 2.0 (Lin, Jenkins, and Abney 2017), this material is virtually absent. In this “major update” and “second generation” of Robot Ethics, what had been marginal in the first version—“marginal” insofar as it was a distant possibility situated at the end of the text—is now effectively pushed out of the book altogether.⁶ This remarkable absence—the lack of something that had been present—is only able to be identified through its trace (Derrida 1982, 65), that is, the withdrawal of something that had previously been there, such that its lack of presence now leaves a mark or trace of its removal.⁷

1.3.4 Exceptions that Prove the Rule

Finally, there are some notable exceptions. But these exceptions, more often than not, prove the rule. In 1985, Robert A. Freitas Jr. published a rather prescient note in the journal Student Lawyer. Although considered to be one of the earliest documented attempts to raise the “questions of ‘machine rights’ and ‘robot liberation’” (Freitas 1985, 54), this short article has had less than twenty citations in the past thirty-five years. This may be the result of: the perceived status (or lack thereof) of the journal, which is not a major venue for peer-reviewed research, but a magazine published by the student division of the American Bar Association; a product of some confusion concerning the essay’s three different titles ⁸; the fact that the article is not actually an article but an “end note” or epilogue; or simply an issue of timing, insofar as the questions Freitas raises came well in advance of robot ethics or even wide acceptance of computer ethics. In any event, this initial effort to situate the question concerning robot rights has been and remains on the periphery of scholarship, occupying a position that is, quite literally, an endnote.

Another notable exception is Blay Whitby’s “Sometimes it’s Hard to be a Robot: A Call for Action on the Ethics of Abusing Artificial Agents” from 2008. Although Whitby does not take up the question of robot rights (the term “rights” does not appear in his essay), his concern is with the potential for robot abuse.⁹ “This is a call,” Whitby (2008, 326) explains, “for informed debate on the ethical issues raised by the forthcoming widespread use of robots, particularly in domestic settings. Research shows that humans can sometimes become very abusive towards computers and robots particularly when they are seen as human-like and this raises important ethical issues.” Whitby’s objective is modest. He only seeks to open up the issue and initiate debate about the possible ethical consequences of future robot abuse. As such, and by deliberate design, he does not portend to offer any practical solutions or answers, which is something for which he is criticized by Thimbleby (2008). He only seeks to get the conversation started in an effort to begin the task of thinking about these potential problems. But even raising the question—even asking that roboticists, ethicists, and legal scholars entertain the question of robot abuse and begin to devise ethical standards in the interest of dealing with these potential complications—was perceived as being premature and unnecessary. As Thimbleby (2008, 341) writes (rather acerbically):

Whitby has not convinced this reader that robotism is distinct from any other technology-oriented ethics. He has not provided persuasive or provocative case studies, e.g., “it’s not the gun that kills but the user” that arise frequently in other technoethical discussions. This reader does not share his sentimental interpretation of “abuse”; robots are not kittens. Robots are easy to build and mass produce; there are no grounds, at least as presented, to justify Whitby’s sentimental title—“sometimes it’s hard to be a robot …” (Aw, I feel sorry for them already.) I imagine, if robots had any notion of what “hard” meant that they’d say it was easy enough to be a robot … In conclusion, then, it should be clear that we disagree with Whitby: there is no urgency to define robotism or any ethics of robots. But as nothing is certain, we should end on a careful note: there is no urgency until the robots themselves start to ask for it.

Thimbleby, it should be noted, does not simply exclude consideration of robot rights tout court but is careful to recognize that if this arguably absurd idea were at some point in the future to become possible, it could only transpire in terms of the Will Theory of rights. The robots themselves will need to ask for it. Consequently, the difference between Whitby’s and Thimbleby’s positions is one that is ultimately grounded in the two different theories of rights. Whitby mobilizes an Interest Theory insofar as he advocates a consideration of the problems and consequences of robot abuse irrespective of whether the robot is capable (or not) of demanding that it not be abused. Thimbleby deploys a Will Theory approach, arguing that until the robots are able to ask for consideration or respect, we simply do not need to be concerned about it. A similar argument is made by Keith Abney (2012, 40), who asserts an exclusive privileging of the Will Theory as the basis for drawing the following conclusion: “we can safely say that, for the foreseeable future, robots will have no rights.” Finally, even though Whitby’s article has been referenced just under forty times in the literature, its main proposal, namely that we take up the question of robot abuse and consider some level of legal/moral protections for artifacts, is only directly addressed in a small fraction of these subsequently published texts (Riek et al. 2009, Coeckelbergh 2010 and 2011, and Dix 2017).

Another seemingly notable exception is Alain and Jérémy Bensoussan’s Droit des Robots (2015), a book written by two attorneys involved in helping to define the emerging field of robot law. The French word droit/droits is, however, a bit tricky and can be translated as either “law” or “rights.” Droit, in the singular and written either with or without the definite article (le droit), is typically translated as “law,” as in le droit penal or “the penal law.” The plural droits, especially when it is written with the definite article les droits, is typically translated “rights,” as in the UN’s La Déclaration universelle des droits de l'homme [The Universal Declaration of Human Rights]. But there is some equivocation concerning this matter, which makes the choice between “law” and “rights” something that is often dependent on context.

The title of Bensoussan and Bensoussan’s book, for instance, would be translated as “Robot Law” and not “Right of Robots.” And this translation is consistent with and verified by the contents of the text, which endeavors to articulate a new framework for responding to the legal challenges and opportunities of robots. In the course of their argument, however, Bensoussan and Bensoussan introduce a charter, composed of ten articles, that they call La Chartes Des Droits Des Robots. In this case, the word droits could be translated as “laws,” “rights,” or both, because the contents of the charter concern not only legal responsibilities for the manufacture, deployment, and use of robots, but also stipulate the legal status of robotic artifacts: “Article 2—Robot Person: A robot is an artificial entity endowed with its own legal personality: the robot personality. The robot has a name, an identification number, a state and a legal representative, which may be a natural person or a legal entity” (Bensoussan and Bensoussan 2015, paragraph 219; my translation). With this second article, Bensoussan and Bensoussan make the case for recognizing robots as a new kind of legal person, like a corporation. This is necessary, they argue, to contend with the unique social and legal challenges that robots present to current judicial systems. In other words, and unlike the vast majority of Anglophone researchers, scholars, and critics, Bensoussan and Bensoussan’s work recognizes the difficulty of formulating laws [droits] that apply to robots without giving some consideration to the legal standing and rights [les droits] of the entity to which these laws would apply. And the one legal right that they believe is paramount for robots is the right to privacy.¹⁰ This is stipulated in the third article: “Personal data stored by a robot are subject to the Data Protection and Freedom regulations. A robot has the right to the respect of its dignity limited to the personal data conserved” (Bensoussan and Bensoussan 2015, paragraph 219; my translation).

It should be noted, however, that the motivation and objective for extending legal protections to the robot remains focused on the human user. Here is how Alain Bensoussan explained it in an interview from 2014:

There is the question of duties, but also of rights. Such as the right to privacy. Indeed, the robot, in its interaction with the elderly or robots who “work” with autistic children, acquires information about their health and privacy. For example, the robot is able to tell a person with Alzheimer's disease “This is going to be your grandson’s birthday” or to an autistic child “Your brother has arrived,” or to an elderly person “Your granddaughter is here.” And this intimacy must be protected by protecting the memory of the robot. (Bensoussan 2014, 13; my translation)

Consequently, Bensoussan and Bensoussan (2015) recognize the importance of explicitly stipulating the social and legal status of the robot and entertain questions regarding both robot law [droit des robots] and robot rights [les droits des robots], or at least one particular claim, the right to privacy. But for all its promise to think and deal with the question of robot rights, Bensoussan and Bensoussan’s Droit des Robots is still restricted to an anthropocentric focus and frame of reference. The one right that the robot would have and would need to have—the right to privacy—is provided not for the sake of the robot but to protect the human beings involved with the robot. It is, therefore, a kind of indirect and instrumentalist way of thinking of the rights of others. Consequently, a robot’s rights are not unthinkable, but their consideration is tightly constrained by, and limited to, the rights that belong to its human users.

1.4 Summary

Investigating the opportunities and challenges of robot rights is already complicated by both linguistic and conceptual difficulties. Linguistically, the term “robot” admits of a wide range of different and not necessarily compatible definitions, characterizations, and understandings. Like the word “animal,” which Derrida (2008, 32) finds both compelling and astonishing—“Animal, what a word!”—the word “robot” is also a rather indeterminate and slippery signifier for a set of ideas and concepts that differ across texts, research projects, and investigative efforts. It is, therefore, and already from the beginning, “a question of words” and “an exploration of language” (Derrida 2008, 33). Even the most general and seemingly uncontroversial of characterizations—the sense-think-act paradigm—remains insufficient to capture and/or represent the full range and complexity that is already available in the current literature. And this “literature” includes not just academic, technical, and commercial publications but decades of science fiction as presented in print (i.e., short stories, novels, novellas, and comic books), stage plays and performances, and other forms of media (i.e., radio, cinema, and television); which in this particular case has determined many aspects of this thing called “robot”—or perhaps more accurately stated, these diverse and differentiated artifacts that have come to be designated and identified with the word “robot”—well in advance of actual efforts in science and engineering. In response to this, it has been decided not to make the usual definitive decision but to attend to and learn to take responsibility for this polysemia. That is, rather than proceeding by way of the “tried and true” approach to defining terminology and issuing a declaration that takes the form of something like, “In this study, we understand ‘robot’ to be …,” we will endeavor to account for, to tolerate, and even to facilitate this semantic diversity and plurivocity. The word “robot,” then, will function as a kind of semi-autonomous, linguistic artifact that is always and already on the verge of escaping from our oversight and control, such that the task before us is not to limit opportunity in advance of the investigation but to account for the full range of significations that the term “robot” puts into play and makes available.

A different but not necessarily unrelated terminological problem occurs with the word “rights.” Unlike “robot,” the term “rights” does have a rather strict and formal characterization. The Hohfeldian incidents parse the term into four distinct and interrelated elements: privileges, claims, powers, and immunities. For most moral philosophers and jurists, as Wenar (2005, 2015) suggests, these four incidents—either alone or in some combination—provide what is typically considered to be a complete formulation of and account for what we understand by the term “rights.” Where things get complicated is in deciding who or what can or should have a share in these four incidents. Who or what, in other words, can have rights? As we have seen, there are two competing theories for addressing this question: the Interest Theory and the Will Theory. The latter sets a rather high bar, requiring that the subject of a right possess the authority and/or capacity to assert, on its own, a privilege, claim, power, and/or immunity that would need to be respected by others. The former provides for a much lower threshold, making it possible for one to decide rights on behalf of the interests of others.

The choice of theory makes a significant difference. Under the Will Theory, for instance, it would be virtually impossible for an animal or the environment to have any stake in the four incidents, and this insofar as they lack λόγος—that influential ancient Greek word that means “word” and can be translated as “speech,” “language,” “logic,” and “reason”—and the ability to articulate and assert their rights. The Interest Theory, however, proceeds otherwise. It recognizes, following Bentham (2005, 283), that the operative question is not “Can they speak?” but “Can they suffer?” This other way of proceeding ¹¹ would institute an entirely different way of deciding between who or what has rights and who or what does not. But this decision between the two theories is itself already a moral determination, for it already makes a decisive cut in the fabric of being separating who can be a moral subject from what remains a mere object situated outside the proper confines of moral concern. And this proto-moral determination—this “moral decision before moral decision” or this “ethics of ethics”—will be and will remain a subject of the investigation. Ultimately, because a decision concerning these two competing theories of rights remains unresolved and debated, we will need to keep both in play by using a strategy that Žižek calls “the parallax view,” which is a way of perceiving and taking account of the difference that is available in the shift from the one theoretical perspective to the other.¹² What is important, therefore, is not selecting the “right theory” of rights at the beginning and sticking with it. What is important is developing the capability to call out and identify which theory is in play in which particular consideration of rights and how each of these frame distinctly different ways of responding to the question “Can and should robots have rights?”

The terms “robot” and “rights” are already complicated enough, but once you put the two words together, you get something of an allergic reaction. “Robot rights” is for many theorist and practitioners simply unthinkable, meaning that it is either unable to be thought, insofar as the very concept strains against common sense or good scientific reasoning; or is to be purposefully avoided as something that must not be thought—i.e., as a kind of prohibited idea or blasphemy that would open a Pandora’s box of problems and therefore must be suppressed or repressed (to use common psychoanalytical terminology). Whatever the reason, there is something of a deliberate decision and concerted effort not to think—or at least not to take as a serious matter for thinking—the question of robot rights. The very idea is openly mocked as ridiculous and laughable; it is cited for engaging in pointless speculation that has the effect of distracting one from the true and serious work that needs to be done; or the subject is pushed aside and marginalized as a kind of appendix or sidebar that could be of interest in the future, but which (for now at least) does not really need serious attention. This is, however, precisely the reason why we must think the unthinkable—to challenge these declarations, assumptions, and orthodoxies; and this needs to transpire for at least two reasons.

First, any dogmatic declaration of this kind should already make us uneasy and suspicious. The fact that this thought—the mere thought that is made available in the association of the two words “robot” and “rights”—has been declared “unthinkable” should give one pause and trigger critical questions like: Who says? Who gets to decide in advance what we can or cannot think? And, perhaps more revealing, what values and assumptions are being protected through this kind of proscription and prohibition? As Barbara Johnson (1981, xv) reminds us, the task of critical thinking is to read backward from what seems obvious, self-evident, and natural, to show how these things already have a history that determines what becomes possible and what proceeds from them. When an idea, like robot rights, is immediately declared unthinkable, that is an indication of the need for critical philosophy and the difficult but necessary task of confronting and thinking the unthinkable.

Second, challenging exclusions and prohibitions is the proper work of ethics. Ethics operates by making exclusive decisions. It inevitably and unavoidably picks winners and losers, and determines who is inside the moral community and what remains outside or on the periphery (Gunkel 2014 and Coeckelbergh 2012). But our moral theories and practices develop and evolve by challenging these exceptions and limitations. Ethics, therefore, advances by critically questioning its own exclusivity and eventually accommodating many previously excluded or marginalized others—women, people of color, animals, the environment, etc. “In the history of the United States,” Susan Leigh Anderson (2008, 480) has argued, “gradually more and more beings have been granted the same rights that others possessed and we’ve become a more ethical society as a result.” But as Stone (1974, 6) has pointed out, this moral progress is difficult insofar as any extension of rights—any effort to extend moral consideration to previously excluded populations—has always been “a bit unthinkable.” In other words, “each time there is a movement to confer rights onto some new ‘entity,’ the proposal is bound to sound odd or frightening or laughable” (Stone 1974, 7). The task of both moral and legal philosophy, therefore, is not (and cannot be) simply to defend existing orthodoxies by shrinking away in fright in the face of alternatives, trigging angry reactions to the possibility of including others, or mocking such challenges with dismissive laughter. Instead the task is, or at least should be, to stress-test and question the limitations and exclusions of existing moral positions and modes of thinking. Defending orthodoxy is the purview of religion and ideology; critically testing hypotheses and remaining open to revising the way we think about the world in the face of new challenges and opportunities is the task of science. The project of philosophy—and moral philosophy in particular—has always been to ask about what has not been thought before or what has been decried and avoided as unthinkable. Doing so for the previously unthinkable thought of “robot rights” is precisely the objective of the chapters that follow.

Notes

1. A similar point was made during an interview published on the Science Museum website:

Interviewer:

Our view of robots is shaped by books and film—do you think this is helpful or misleading?

Alan Winfield:

Good question! I think robots in science fiction are both helpful and misleading. Helpful because many roboticists, myself included, were inspired by science fiction, and also because SF provides us with some great examples of “thought experiments” for what future robots might perhaps be like—think of the movie AI, or Data from Star Trek. (Of course there are some terrible examples as well!) But robots in science fiction are misleading too. They have created an expectation of what robots are (or should be) like that means that many people are disappointed by real-world robots. This is a great shame because real-world robots are—in many ways—much more exciting than the fantasy robots in the movies. And the misleading impression of robots from SF makes being a roboticist harder, because we sometimes have to explain why robotics has “failed”—which of course it hasn’t! (Winfield 2011b)

5. The response from the general public was no better as Seo-Young Chu (2010, 215) documents by way of user comments:

Most humans would rather dismiss such claims as frivolous and insubstantial than acknowledge them as part of reality. “This is a silly concept,” one commentator remarked in response to the 2006 study released by the British government: “Granting a robot rights would be akin to granting the right front tire on my car rights” (Bryansix). Another commentator expresses his resistance in harsher language: “Robot rights are a joke and so are the morons that even consider them” (GoldFish). A third commentator voices his objection to robot rights in the most violent way possible: “If I buy the robot, I should be able to do whatever the hell I want to it. Including beatings, burnings, disabling, beheading” (Keep Robots Slaves). From amused skepticism to disgust and vehement hostility, such attitudes of mental rejection are symptomatic of the cognitively estranging nature of robot rights.

6. It may, of course, be objected that this interpretation simply makes too much out of what is nothing more than editorial exigency. When I asked Patrick Lin about it (via a personal email from August 2017), he offered two explanations for this perceived difference in the two versions of Robot Ethics: (1) “I wouldn’t read too much into what we call our sections—that doesn’t really track any perceived trend or delimit our areas of interest.” And (2) the “book does engage the question of rights in several chapters … it just made more sense to organize them topically in a different way.” If you track the term “rights” as it occurs in the second version, however, there is just one essay that gives passing consideration to the rights of robots by way of its introductory material (Darling 2017). The remaining contributions, when they do take up and engage the question of rights, focus attention on the rights of human individuals and communities. Even the consideration of robot personhood, which Sparrow (2012) had developed in the first version, is now limited to matters of robot liability (White and Baum 2017, 70).

7. The logic of this operation was already formulated in Plato’s Timaeus. The dialogue begins with a kind of roll call where Socrates accounts for who is present (or not) by way of counting: “One, two, three,—but where, my dear Timaeus, is the fourth of our guests of yesterday, the hosts of today?” (Plato 1981, 17a). In enumerating the individuals present before him, Socrates notices a difference between the number of people that had been present during the previous day’s discussion (presumably the Republic, insofar as the Timeaus is situated as its sequel), and the number of participants who are currently present at the beginning of the current discussion. In the process of comparing what had been yesterday to what is today, Socrates identifies an absence. There is someone missing, someone who was supposed to be there but now is not. This absence, or lack of presence, supervenes and takes place insofar as someone is not there and in their place. Consequently the “missing fourth” of the Timaeus is available and only comes to presence by way of a taking account of this difference. The absence of the one who should have been present only comes to presence by way of this trace of their withdrawal.

9. This attention to “abuse” may be a product of the venue in which Whitby’s essay initially appears, a special issue of Computers and Interaction addressing “the abuse and misuse of social agents.” Although the term “abuse” might give some readers pause, insofar as the word might be perceived as being too anthropomorphic and not applicable to technological objects, it is explained and justified in the editors’ introduction.

We recognize that the title of this special issue, Abuse and Misuse of Social Agents, may annoy or disappoint some readers because of its metaphorical implications. Clearly, the word abuse, when defined as the intent to cause suffering, cannot be applied to inanimate objects, as they are incapable of recognizing, experiencing, or understanding suffering. … Although we agree with the above statement, we do believe that the metaphor of social agency by itself tends to promote technology to a level that cannot escape the possibility of abuse. If it makes sense to say users are being polite in their responses to a conversational agent when obeying the social rules of politeness, then why would it not make equal sense to say a user is being abusive in his or her responses when using language that, if applied to people, would be considered abusive? (Brahnam and De Angeli 2008, 288)

11. The issue of suffering has the effect, Derrida (2008, 27) argues, of “changing the very form of the question regarding the animal”: “Thus the question will not be to know whether animals are of the type zoon logon echon [ζωον λόγον έχον], whether they can speak or reason thanks to that capacity or that attribute of the logos [λόγος], the can-have of the logos, the aptitude for the logos (and logocentrism is first of all a thesis regarding the animal, the animal deprived of the logos, deprived of the can-have-the-logos: this is the thesis, position, or presumption maintained from Aristotle to Heidegger, from Descartes to Kant, Levinas, and Lacan). The first and decisive question would be rather to know whether animals can suffer” (Derrida 2008, 27). For more on this shift in the question concerning the moral standing of others, see the final chapter and Gunkel (2012).

1
Thinking the Unthinkable

1.1 Robot

1.1.1 Science Fiction

1.1.2 Indeterminate Determinations

1.1.3 Moving Target

1.1.4 Results/Summary

1.2 Rights

1.2.1 Definition

First Order Incidents

Second Order Incidents

1.2.2 Theories of Rights