Michael Littman spoke at TEDxProvidence September 30th 2017 on “How I learned to stop worrying and be realistic about AI”
Michael Littman spoke at TEDxProvidence September 30th 2017 on “How I learned to stop worrying and be realistic about AI”
Because robots’ decision-making abilities are not at the level of humans’, there are liabilities that come with industrial robots, automated vehicles, caretakers, and other positions that involve life-or-death situations. More so than robots rebelling and taking over the world, people should be worrying about robots malfunctioning or falling into the hands of the wrong people.
There have been three reported incidents of industrial robots causing deaths in factories, the latter two of which involved human error at least in part. In 1979, a Michigan Ford Motor robot’s arm crushed a worker while both were gathering supplies. The jury awarded the employee’s family $10 million, the state’s largest personal injury award ever at the time. Ford Motors was blamed for the incident because of a lack of safety precautions, including an alarm that should have sounded when the robot approached.
In 1984, a die-cast operator was pinned between a hydraulic robot and a safety pole. The worker was blamed for entering a robot envelope, which was prohibited in training. The company subsequently installed a fence to keep unauthorized workers away from the robots. This incident led the National Institute for Occupational Safety and Health’s Division of Safety Research to make recommendations in “ergonomic design, training, and supervision” for industrial robots, including design measures for the robot envelope, safety precautions for workers and programmers, and instructions for training. Supervisors were advised to emphasize in training that workers should not assume that a robot will keep doing its current activity or stay still when stopped.
More recently, in 2009, a robot’s arm again crushed and killed a worker at a Golden State Foods bakery in California. According to the inspection detail, the case is not closed yet, but it appears that Golden State Foods has to pay several fines, over $200,000 total. The incident was chalked up partially to the worker’s lack of precaution. The inspection report reads: “At approximately 7:55 a.m. on July 21, 2009, Employee #1 was operating a robotic palletizer for Golden State Foods, Inc., a food processor and packager for fast food restaurants. She entered the caged robotic palletizer cell while the robotic palletizer was running. She had not deenergized the equipment. Her torso was crushed by the arms of the robotic palletizer as it attempted to pick up boxes on the roller conveyor. She was killed.”
The International Organization for Standardization (ISO) has published 10 standards for industrial robots and drafted two standards for personal care robots. The American National Standard and the Robotics Industries Association have jointly developed detailed safety regulations for industrial robots, which were recently updated to incorporate the ISO’s guidelines.
A newer development for governments to respond to is the automated car. Self-driving cars are legal in California, Nevada, and Florida for testing on roads as long as there is a human behind the wheel. The only Google car accident so far occurred when the car was controlled by a human, as opposed to a computer. When asked who would get ticketed when the Google car ran a red light, co-creator Sergey Brin responded, “self-driving cars do not run red lights.”
Currently, however, the United States National Highway Traffic Safety Administration (NHTSA) is working on a set of rules governing the use of self-driving cars. The recommendations include special training to obtain a license for an automatic car and a requirement that those testing the vehicles report all accidents. The NHTSA proposal points out that most cars already have automated features, such as brake pulses during potential skids, automatic braking during potential collisions, and cruise control.
Another liability for robots is hacking. People can hack into not only computers, but also any machine that is part of a network— including cars with the features described above. To illustrate this possibility, computer scientists Charlie Miller and Chris Valasek hacked into a Ford Escape and caused it to crash even when the driver hit the brakes. Robot hacking similarly contains the potential for physical damage.
Increasingly automated machines are also bringing up security questions.1 Some medical and domestic robots record personal information and behavioral patterns, which privacy laws do not yet address. If information is not kept between the machine and its user, the consequences for medical robots, automated vehicles, or smart houses could be dire. Apps such as Nest, which connect your phone to your house in order to control the temperature, could give a hacker information about your home — and if it were a robot the app was controlling, much more.
Current hacking laws are difficult to apply to the 21st century, let alone beyond. The Computer Fraud and Abuse Act, passed in 1986, has been used to prosecute internet users for finding loopholes in websites without even revealing the information they find.
When we ascribe blame for a crime, we usually ascribe it to an individual who has acted maliciously or carelessly. But those words don’t really apply to robots, who act based on programming, intended or not. Adapting our laws to robots may require us to rethink agency, or at least to think more about who the agents are in these situations.
1 A Roadmap for U.S. Robotics: From Internet to Robotics. 2013 Edition. March 20, 2013.
“I will remember that artificially intelligent machines are for the benefit of humanity and will strive to contribute to the human race through my creations.”
The field of robot ethics often focuses on ethical problems with creating robots. But because certain populations are in need of the extra care, protection, or labor that robots provide, it can be unethical not to create robots. The US Congress’s Robotics Caucus Advisory Committee’s 2013 Roadmap for US Robotics documents how robots can improve and even save lives. At work, they can “assist people with dirty, dull, and dangerous tasks;” at home, they can offer “domestic support to improve quality of life;” and in warzones, “robots have already proven their value in removing first-responders and soldiers from immediate danger.” The US government’s National Robotics Initiative (NRI) hopes to encourage robotics in order to boost American manufacturing, space exploration, medical discoveries, and food safety.
Many roboticists share the vision of using robots to solve some major problems the world currently faces. For example, an area receiving particular attention among researchers is elderly care. The number of people over 80 worldwide is expected to double by 2030, leading to increased demands for living assistance. Smart homes and robot caretakers can help these people stay independent longer (1). Even in military use, robots are not necessarily killing machines. iRobot CEO Colin Angle said that his PackBot has saved the lives of EOD technicians by hunting IEDs in Afghanistan and Iraq. Furthermore, drones have proven applicable beyond the military: Association of Unmanned Vehicles International CEO Michael Toscano estimated in July 2013 that agriculture would overtake defense as the most common use for drones in the next 10 years, noting that this advancement will help feed an ever-growing population.
Those who fund and use robots can ensure that robots’ potential to do good for the world does not go to waste. An active approach may be necessary because research may otherwise be driven simply by potential profits.
“A profitable market for expensive new technology is not of itself sufficient moral defense for allowing widespread sales.” -Whitby (2012)
The first ethical question in the development of robots is what kinds of robots to make. Are we creating the kinds of robots that our society needs? The answer to this question depends on the motivations not only of researchers but also of their sponsors. The main sources for robotics funding are the government, businesses, and research institutions.
The government has played a particularly significant role in robotics funding since 2011, with the advent of the NRI, whose mission is to fund research that facilitates human-robot interactions. Its 2011 and 2012 solicitations were subtitled “The realization of co-robots acting in direct support of individuals and groups” and stated that proposals would be selected by a peer-review process. The National Science Foundation listed the recipients of $30 million of the $50 million awarded for this initiative (the destination of the other $20 million is unclear). The plurality of the projects listed (12 out of 31) named medicine as an application for its robots, including aid in rehabilitation and surgery, followed by domestic care — either general service robots or those geared toward the elderly or incapacitated — and manufacturing, such as industrial robots for use in factories (2).
As part of the NRI, the Defense Advanced Research Projects Agency (DARPA)’s FY2012 solicitation for the Defense University Research Instrumentation Program (DURIP) encouraged proposals for robotics research equipment applicable the military (Army Research Office). Of the 190 recipients of DURIP grants, which totaled about $55 million, 22 projects involved robots, autonomous machines or vehicles, or artificial intelligence. Many robotics-related projects were not directly or solely applicable to the military, including 3 grants for general human-robot interaction research and 7 for ocean surveillance and research. Others objects of study included the response of miniature air drones to wind and water-splitting catalysts for autonomous energy generation.
DARPA will account for nearly half of the government’s Research and Development funding under Obama’s FY2014 plan. Though this suggests that research applicable to the military may have priority, it does not necessarily mean that the robotics projects DARPA sponsors are primarily for killing. For example, the agency is currently sponsoring a $2 million “Robotics Challenge” in which robots compete in a series of tasks simulating natural and man-made disasters. The press release cites robots that diffuse explosives as an example of the type of innovations DARPA is looking for. The Robotics Challenge has sponsored robots made by universities, companies, and government organizations such as NASA.
However, government funding of robotics has its critics, perhaps the most outspoken being Oklahoma Senator Tom Coburn. Coburn criticized a $1.2 million NSF grant to a laundry-folding robot at UC Berkeley and a $6000 grant for a “Robot Rodeo and Hoedown” at a computer science education symposium to introduce teachers to robotics (3). A researcher in the UC Berkeley group countered that the laundry-folding algorithm was a big step for robotics more broadly, and the organizers of the robot rodeo event explained that they were trying to get more students interested in computer science.
Venture Capital funding is also a major source of funding for robotics. In 2012, the 22 most funded robotics companies accrued almost $200 million total from VC funds (4). Like the NRI funds, the plurality went toward medicine — about $84 million of the money awarded by VC funds went toward companies that create robots for surgery (e.g. MedRobotics and Mazor Robotics), bionics (e.g. iWalk), and hospital service (e.g. Aethon). Another $30 million went to Rethink Robotics for the manufacturing robot Baxter.
Some have criticized excessive funding for entertainment robots, such as a robotic toy car largely funded by Andreessen Horowitz. The car’s maker, the new start-up Anki, raised $50 million total in 2012 and 2013. Anki’s founders believe this investment is justified because their technology can lead to more advanced and affordable robots down the line.
Universities themselves sometimes fund robotics research, but these funds are often funneled from the sources above — in the United States, particularly the government. The Robotics Institute at Carnegie Mellon, the first and largest robotics department at any US university, lists its biggest sponsors as the Department of Defense, DARPA, the National Aeronautics and Space Administration, the National Institutes of Health, and the NSF. The Georgia Tech Center for Robotics and Intelligent Machines and the MIT Media Lab get sponsorship from companies in the robotics industry in exchange for developing their products. This model appears to be the most common one in Japan, China, and South Korea, according to a WTEC report. Google and other companies have grants specifically for funding technological research, along with private foundations such as the Keck Foundation and the Alfred P. Sloan Foundation.
While grants are usually given for specific projects, some departments retain control over which projects their funds go toward. The Robotics Institute at Carnegie Mellon has a budget of $65 million per year, which it has used to fund 30 start-up companies in addition to supporting its own institution. Carnegie Mellon also, along with companies like Google and Yahoo and private donors, sponsors a foundation called TechBridgeWorld that has aided technological progress in developing areas of the world by funding innovations such as an “automated tutor” to improve literacy. This exemplifies how academics and researchers can use their knowledge of what people need and how to most efficiently meet those needs to influence production.
Because robotics companies, like any company, strive to make money, the purchase of robots also affects which robots get made. The International Federation of Robotics estimated that 2.5 million personal and domestic robots (e.g., those that do household chores or assist people with medical problems) and about 16,000 professional robots for use in various workplaces were sold worldwide in 2011. The professional robots consisted mostly of military (40%), and agricultural (31%) robots, but also of robots for other uses like medicine (6%) and manufacturing (13%). The Robotics Industries Association estimated during July 2013 that 230,000 robots were currently in use in United States factories, most often in the automotive industry.
Fifty years ago, we couldn’t imagine computers as part of our everyday lives. Now, we can’t imagine our lives without them. Robots are the new computers, in that they are capable of revolutionizing our society and economy. Whether this revolution is for better or for worse is up to all the aforementioned players.
One of people’s biggest concerns regarding the possibility of owning robots is losing control of these robots(1). Getting robots to cooperate with humans is a challenge given the numerosity and complexity of the rules governing social conduct. Isaac Asimov illustrated this difficulty in his short story “Runaround,” in which the rules governing a robot’s behavior conflict, cancel one another out, and supersede one another in unpredictable ways. Whether instilling appropriate behavior in robots is the job of its designers or its owners, and whether these behaviors should be intrinsic or learned, is up for debate. What is clear is that we need a better understanding of what guides moral behavior in humans as well as how to program these codes into robots to ensure ethicality and effectiveness.
Advanced social robots may need built-in ethical guidelines for their behavior in order to avoid situations in which robots are used to manipulate or harm people (2). The first step is getting people to agree on what these guidelines are — for example, should robots merely avoid bumping into people, or should they politely evade them by saying “Excuse me”? (3). The next step is implementing these guidelines. Wendell Wallach and Colin Allen’s book Moral Machines: Teaching Robots Right from Wrong describes two approaches to robot morality: programming the robot to predict the likelihood of various consequences of various actions, or having the robot learn from experience and acquire moral capabilities from more general intelligence.
But as it stands, the social abilities of robots are currently very rudimentary (4). According to the 2013 Roadmap for U.S. Robotics, “medical and healthcare robots need to understand their user’s state and behavior to respond appropriately,” which requires a multiplicity of sensory information that current robots have trouble collecting, let alone integrating into something meaningful and utilizing to inform their own behavior. As an illustration, voice recognition features on cell phones often have trouble understanding what the user is saying, and security cameras often fail to recognize human faces (3). The roadmap also mentions the importance of empathy in healthcare, rendering socially assistive robotics (SAR) for mental health problematic. Joseph Weizenbaum, creator of the first automated psychotherapist in the 60s, balked at the idea that his creation, Eliza, could replace a human therapist (5). Today, there are more advanced automated therapist programs, but like Eliza, they mostly echo what the user says and ask generic questions. Therapeutic robots have been more effective in the realm of interactive toys for autistic children, as long as these are viewed as toys and, like any robotic invention, not as replacements for human relationships.
1 Ray, C., Mondada, F., & Siegwart, R. (2008) What do people expect from robots? IEEE/RSJ International Conference on Intelligent Robots and Systems,3816–3821.
2 Scheutz, M. The Inherent Dangers of Unidirectional Emotional Bonds between Humans and Social Robots. Robot ethics: the ethical and social implications of robotics. Ed. Lin, Patrick. Cambridge, Mass.: MIT Press, 2012.
3 Murphy, R. R., & Woods, D. D. (2009). Beyond Asimov: The Three Laws of Responsible Robotics. IEEE Intelligent Systems, 24(4), 14–20.
4 “A Roadmap for U.S. Robotics: From Internet to Robotics.” 2013 Edition. March 20, 2013.
5 Schanze, J. (Director) (2011). Plug & Pray [Documentary]: B.P.I.
Children get attached to their toys. They bring them on trips, care for them, ascribe human qualities to them, and even worry about them. As it turns out, adults do too, especially when these toys lend themselves to human empathy. An object triggers an empathetic response when it moves on its own, has eyes, and/or responds to stimuli in the environment. But it doesn’t need to have all these qualities. Two-thirds of Roomba owners have named their automated vacuum cleaner (1). If you’ve ever seen a Roomba, you know it basically just looks like a big disk with some buttons.
Some designers and manufacturers aim for realism in humanoid robots, even believing that these robots will be so realistic as to be sentient. David Hanson, known for extremely lifelike and seemingly emotional robots such as Jules, writes on his website that his robots have “the spark of soul—enabling our robots to think, feel, to build relationships with people […].” Cynthia Breazeal, a robot engineer at MIT, said that after watching science fiction films as a child “it was just kind of a given for me that robots would have emotions and feelings.” Perhaps because this is interesting to consumers the same way a very exact painting would be, the advantage of this approach is often assumed rather than examined.
Scientists who aim to make robots relatable often justify their approach by pointing out the advantages of people liking and feeling cared for by their robots. Breazeal, creator of cutest-robot-ever Leonardo, said she aimed to design a robot “that understands and interacts and treats people as people.” She believes that robots put to tasks such as household chores and childcare should be able to socialize and respond to humans, just like other humans, so that people can interact with them more easily. The Roomba, she points out, does not recognize people, so it runs into them all the time.
Making machines that treat humans as humans, however, is a slightly different task than making machines that demand to be treated as humans themselves. Wendell Wallach and Colin Allen have been thinking about how to create “moral machines.” For them, this means bestowing robots with the ability to read human emotions and understand human ethical codes. And sometimes it means expressing empathy for people. A semblance of empathy is especially important for therapeutic robots (2).
But while most people want well-behaved robots, some argue that we should not want deceptively human ones. Matthias Scheutz at Tufts is an outspoken critic of social robots. When robots are convincingly human, they may be not only endearing but also persuasive to humans. Humans may become excessively dependent on and devoted to their robot companions, to the point of feeling and behaving as if these robots are sentient. This gives robots great power over humans, which is especially dangerous if their intentions are not honorable (1).
Science and science fiction alike have acknowledged the complications of human-robot relationships. In the 1991 novel He, She and It by Marge Piercy, a human falls in love with a robot who (spoiler alert) ultimately must self-destruct in battle to fulfill the commands of the people who created him. Renowned roboticist Joseph Weizenbaum explained in the documentary Plug and Pray that he believes love between humans and robots would ultimately be illusory because a robot cannot understand what it is like to have the personal history of a human.
As such a future grows ever nearer, we will have to draw the line between developing robots with the social skills to carry out the desired tasks and creating a deceptive appearance of conscious thought.
1 Scheutz, Matthias. “The Inherent Dangers of Unidirectional Emotional Bonds between Humans and Social Robots.” Robot ethics: the ethical and social implications of robotics. Ed. Lin, Patrick. Cambridge, Mass.: MIT Press, 2012.
2 “A Roadmap for U.S. Robotics: From Internet to Robotics.” 2013 Edition. March 20, 2013. http://robotics-vo.us/sites/default/files/2013%20Robotics%20Roadmap-rs.pdf
“We already have impressive models and simulations of a couple dozen of the brain’s several hundred regions,” writes Ray Kurzweil in The Singularity is Near.1 But how accurate can these models get before they stop being models, and how advanced can artificial intelligence be before it is no longer artificial?
We are getting closer and closer to creating machines that do all the work of human brains — work that previously required thoughts and feelings. David Hanson’s robot Jules, for instance, mimics not only awareness but also self-awareness when he muses over his identity as a robot in this video. It’s hard not to feel sympathetic to his human-like struggles. Hanson’s website implies that he thinks his robots are or will be sentient: “The Character Engine A.I. software adds the spark of soul—enabling our robots to think, feel, to build relationships with people as they understand your speech, see your face, hold natural conversations, and evolve.”
Though there are robots that realistically portray such mental states, we usually view them as just that: portrayals. The problem is, we are unprepared to identify when these semblances of thoughts and feelings turn into actual thinking and feeling.
Our knee-jerk reaction is still that such robots aren’t at the level of humans. But who says you have to be human to be sentient? Many ascribe consciousness to simpler animals. And many machines have more sophisticated circuitry than many animals. One might argue that animals are biological and thus not a valid comparison, but we don’t know that the stuff of consciousness is biological. In fact, we don’t know what the stuff of consciousness is at all because there is no way of knowing which beings are conscious. We could ask them, but they could lie. We could probe their brains (or equivalent internal circuitry), but there is no way of knowing which objective data indicates subjective experience (though we can make reasonable presumptions based on which brain areas seem correlated with consciousness in humans; see my conclusion below).
Kurzweil suggests that we use the same guidelines we use for humans: We don’t know other people are conscious — called “the problem of other minds” in philosophy — but we give them the benefit of the doubt because of their apparent emotions and social abilities. Many give animals the benefit of the doubt too for the same reason, and because they would rather unnecessarily protect creatures that don’t suffer than harm ones that do. The American Society for the Prevention of Cruelty to Robots, started in 1999, offers a similar solution: “the question of self-awareness comes down to an individual leap of faith: ‘I am self-aware, and I see others behaving in similar ways, so I will assume that they are self-aware, also.'”
The loophole I see in these arguments, aside from the fact that attributions of consciousness are necessarily assumptions, is that humans are predisposed to ascribe sentience to beings with certain appearances and behaviors — specifically, those which are human-like. People generally ascribe minds to things that move on their own, have eyes, and react to stimuli.
Terry Bisson critiques biases in other-mind attribution in the short story “They’re made out of meat,” in which (spoiler alert) conscious planetary gases debate whether humans could possibly think, given that they’re just “made out of meat.” The point is, anything could be conscious, but we assume that because we are and this appears to come from our brains, all conscious beings must have (organic) brains. If the possibility of conscious hydrogen seems too crazy, the possibility of consciousness in robots that socialize and express understanding of humans is far less outlandish.
On the flip side of our anthropocentric affinity for beings resembling us, people may easily attribute consciousness to things that aren’t really conscious, even when they know intellectually what sort of being they’re dealing with. For example, people report naming their roombas and feeling guilty that they make them do all the work.2 This can cause psychological trauma when people ascribe minds to machines destined for destruction, such as military drones and IED attackers.2 I remember my 7-year-old self lying in bed at night, worrying about the well-being of my Tamagotchi.
We need guidelines for the ascription of sentience that are more reliable than human instincts. Though in their infancy and incapable of certainty, elaborate theories of consciousness are emerging to help us decide when moral restrictions apply to other creatures. Physicist Roger Penrose has proposed criteria for consciousness that are implementable in organisms and computers, but, he argues, not in their current state: They would need a different mode of causation that follows a different form of physics.
It’s very hard to think of current robots as conscious. But eventually (and possibly soon!), we will have to make a leap of faith in believing in robot minds, like we do with human minds or animal minds. Though consciousness is a subjective experience, I believe we can find reasonable, objective ways for trained professionals to make educated guesses. True, for all we know our pillows hurt when we punch them, but this is a philosophical question far broader than what is necessary to create conscious robots.
For now, this is the safest approach I can see: If sentience seems like a possibility, treat your machines with respect. This assumption prevents harm to something that could be sentient, keeps our consciences clean, and sets an example for the benevolent treatment of living things and property alike — even if we don’t know which we’re dealing with.
1 Kurzweil, Ray. The Singularity is Near: When Humans Transcend Biology. New York: Viking, 2005.
2 Scheutz, Matthias. “The Inherent Dangers of Unidirectional Emotional Bonds between Humans and Social Robots.” Robot ethics: the ethical and social implications of robotics. Ed. Lin, Patrick. Cambridge, Mass.: MIT Press, 2012. Print.
You may have seen recent Kia Forte commercials featuring “Hotbots” — their version of a female sex robot commonly known as a fembot.1 In this ad, it is implied that the owner of the car is using the robot for sex. Kia used a human actress for the robot,2 but such advanced technology is not far off, and robots* for such uses already exist, ranging in complexity from inflatable dolls to realistic-looking and -feeling machines with built-in motion. They are especially popular in Japan and South Korea, where individuals rarely own them because they sell at 6-figure prices**, but rent-a-doll “escort” services make a profit by renting them out to customers.3 There is also a significant following among American men with “robot fetishes” who use fembots as sexual and/or romantic partners.4
Personally, I find the whole practice creepy. But as with any sexual practice, finding something aversive is not a reason to oppose others doing it. In my opinion, though, there are some justifiable reasons one might oppose (if not prohibit) sexbots:
They can promote unhealthy attitudes toward relationships. Those who market sex robots aim to produce, or at the very least encourage, desires for unbalanced relationships in which one has total control over one’s partner (particularly, as it stands now, where men have total control over manufactured women). Teaching people that sex with an unresponsive, undesiring partner is desirable could perpetuate rape culture. One male user raved that the dynamic was “as close to human slavery as you can get.”4
They could be considered prostitutes. South Korean authorities have debated whether rent-a-doll escort services, which were created to dodge prostitution laws, should also be illegal.3
They could encourage gender (and possibly any type of) stereotypes. Manufacturers already ascribe gendered characteristics to other types of robots.5 Sex robots open the door for the promotion of stereotypes of women as submissive fulfillers of men’s desires or “men as perpetually-ready sexual performance machines.”6 A current line of sex robots has different models representing some problematic stereotypes, such as Wild Wendy and Frigid Farrah (racializing as well as gendering robots).
They could promote misogyny. Fembots are significantly more popular than their malebot counterparts, and they possess traditional “feminine” qualities: submissiveness, adhesion to beauty ideals, and desire (if one could call it that) to please men without any desires of her own. Like the Stepford Wives, fembots allow men to fulfill patriarchal fantasies in which women exist solely to please them. One male user said “the whole idea of man taking charge” attracted him to fembots.4
They may replace real relationships and distance users from their partners or other potential partners. A plurality of Americans polled indicated that sex with a robot while in a relationship is cheating.5 Whether or not it is infidelity, robots designed to meet needs that humans cannot, such as “lack of complaints or constraints,”3 may give humans unrealistic standards to live up to in order to capture their partners’ attention. In addition, having sexual and possibly even romantic interactions that don’t require any effort may discourage users from meeting real people. In addition, people with partners who are “immune .. to their clients’ psychological inadequacies”3 may not address these inadequacies, making it even harder to form real relationships.
Users could develop unhealthy attachments to robots. Unless robots achieve consciousness (which merits a separate blog post), people could develop feelings for robots who don’t feel anything back. This may sound silly, but people already name and dress up their roombas. It is natural for people to develop attachments to sexual partners — whether this applies when the partner is not human remains to be seen — and some even hope for their sexbots to double as romantic partners.4 An asymmetrical relationship of this nature could end up unfulfilling, isolating, or depressing.
Whether sexbots are ethical is a separate issue from whether they should be legal. For countries and US states that ban sex toys altogether, the decision will be obvious. Countries that ban prostitution, like South Korea, may debate whether a machine can be considered a prostitute. Some laws banning sex toys in the US have been struck down on the grounds of privacy, and I suspect these same arguments will come up for sexbots. David Levy, a scholar on the topic, argues that sexbots are no more legally problematic than vibrators.3 But sexbots have raised and will further raise ethical, if not legal, issues specific to their goal of simulating human beings.
* It is debatable whether the current generation of sex dolls can be considered robots, but they do seem to be headed in that direction.
**This information may be outdated, as there are currently sex robots on the market for 4-figure prices.
1 “Kia Chalks up Another Ad as a Sexist Fail | About-Face.” About-Face. Accessed June 18, 2013. http://www.about-face.org/kia-chalks-up-another-ad-as-a-sexist-fail/.
2 “2014 Kia Forte Super Bowl Ad Features Sexy Robots | Edmunds.com.” Edmunds. Accessed June 18, 2013. http://www.edmunds.com/car-news/2014-kia-forte-super-bowl-ad-features-sexy-robots-disrespectful-reporter.html.
3 Levy, David. “The Ethics of Robot Prostitutes.” Robot ethics: the ethical and social implications of robotics. Ed. Lin, Patrick. Cambridge, Mass.: MIT Press, 2012. Print.
4 “Discovery Health ‘Sex Robot Roxy: Sex Robot’.” Discovery Fit and Health. Accessed June 18, 2013. http://health.discovery.com/tv-shows/specials/videos/sex-robot-sex-robot.htm.
5 “Robot Sex Poll Reveals Americans’ Attitudes About Robotic Lovers, Servants, Soldiers.” Huffington Post, April 10, 2013. http://www.huffingtonpost.com/2013/04/10/robot-sex-poll-americans-robotic-lovers-servants-soldiers_n_3037918.html.
6. Brod, Harry. “Pornography and the alienation of male sexuality.” Social Theory and Practice 14.3 (1988): 265-84.