Do We Need Robot Lovers?

Posted on by
Reply

Hello Internet world. Welcome to Blog #2! In this blog I want to discuss the issue of robot love as a goal of researchers and software engineers alike. Many of the arguments for and against this goal have been discussed by Blay Whitby and his article on the status of Robot lovers[1]. In his work, Blay Witby discusses the motives of creating robot lovers. The claim he cites goes as follows. People may want to create robot lovers to test the bounds of human robot interaction, and love may naturally flow from the organic relationship that comes from it. People may also want to create robot lovers because they may be better and more obedient than human lovers. The third reason is the most pressing and the most interesting. A third reason he cites for creating robot lovers, is to benefit the socially isolated, who cannot or will not involve themselves in loving human relationships. I believe that the third claim for making robot lovers is ethically wrong. We should not sponsor technology in ways that substitute human interaction with artificial care. Blay Witby discusses this problem slightly, but I would like to go more in depth, by arguing that we should not make robot lovers for the point of contention above (mainly to serve people who are socially isolated). I base myself on the preposition that human social relationships are necessary for a quality life. Humans, by their very nature, are social animals. I will thus begin by arguing that human interaction is ethically good. The need to be social can be discussed as a base human desire, and so it would be (for the sake of this argument) similar to the base need of nutrition to fight off starvation. We do not have the technology to completely fill the need for human social conduct with robots. We do however; have the capabilities to fix the problems by other means. Socially isolated people can be brought into society in non intrusive ways. There are programs and help groups already in place. Focusing on the ways in which humans can be helped by robots, seems to mask the problem, which already has a definite solution. We should not make robots for the socially isolated because we are substituting a known fix for a potential fix, which is neither biological nor feasible right now. If more effort and funding goes into a potential fix than a real fix, less people right now will be socialised. Socialising is naturally good, as stated above. Blay Whitby also identifies another concern, and I believe that it is something worth addressing in this blog. In Blay’s work, he discusses quite briefly the problem with capitalist motivations in regards to the human need to socialise. Capitalist motivation would see more potential for profit in a world that would prioritise advances in robot/human interaction over a world that prioritises things like help groups. In closing, I believe that emphasis should be put on humans helping humans socialise, rather than putting the effort and the priority on finding alternative solutions.

 

I have embedded a link which shows that efforts have already been made to get robots to seem more social. The emotional responses of the robot vary, but the intent is to have robots interact in more appropriate ways. If corporations want to have robots substitute humans in inter human relations, this may be a good start.

Later Robot Lovers!

[1] Whitby, Blay Do You Want a Robot Lover  239.

Future Warriors and Our Responsibility

Posted on by
Reply

PW Singer on Robots of War

It is an ever-growing situation that will likely garner argumentation from politicians and historians for years to come. This emerging industry of military robots has become an important discussion for our generation. Robots are a part of our society and our military now and will only continue to gain more ground in these respects. The fact that the robots described in this video, and those seen in class and in pictures on the internet are merely considered the “Model T” of robots implies not only that robots are not going away any time soon but in fact are going to grow and become an industry staple within the military as they continue to become more advanced and autonomous. As PW Singer states in his presentation, robots will become the warrior of the future sooner than one might think, especially if Moore’s Law holds true.

The future of war thus has some significant social and ethical concerns that have become ripe for discussion in our generation.  It is a discussion that at times is difficult to garner attention because it seems so futuristic, like something that we see in a movie, but the numbers of robots that are presently being used in war efforts show the reality of this situation. I do not intend to suggest or even rhetorically imply that this is necessarily a bad thing, in fact I think it to be a positive advancement for our war efforts present and future. The ethical concerns are quite apparent for these military industrial advancements and are worth discussing to begin with, but what I would like to focus on is the implications going forward in terms of motivation for war.

Singer spoke about the EOD units and the lives that Packbots save on a regular basis. This implies that robots are being used to save human lives, which is likely the reason they were put into the army in the first place. With the invention of the Improvised Explosive Device, which is likely the most dangerous weapon to land units in the military, robots have been the answer to the IED problem. It is much easier to send a piece of $5000 metal into the battlefield to scope out an IED than it is to send a man who has a wife and family at home in to find that same IED.  Without a doubt this example proves the useful capabilities of robots within warfare. Yet with rapid expansion comes even greater changes.

The Packbot is likely the most peaceful and ethically permissible bots in military use, as it is primarily used to prevent soldiers from being killed by IEDs. What about the other bots though, such as SWORDS a robot capable of carrying pretty much any weapon under 300 pounds; the likes of which include .50 calibre machine guns, 40mm grenade launchers, and anti-tank rocket launchers (Singer, 2009). This robot is quite obviously created to kill and/or destroy things that it comes into contact with. It should be noted that the SWORDS is however controlled by a human and has a much greater sight capability and effectiveness when shooting in comparison to its human counterpart that has to compensate for such things as an elevated heart rate and heavy breathing due to stress or anxiety in a war setting (Singer, 2009). The use of SWORDS and the classification of it being one of the “Model T” robots of this generation, implies that it will likely gain greater amounts of autonomy going forward. The implication of military robots gaining a greater autonomy in the near future is a change in how we understand war-crimes.

Who is to be held responsible when these autonomous bots go into a war-zone and kill hundreds, or even thousands of people? The likely answer is that these robots are likely to be better prepared for the war than humans, due to the fact that they have no emotions that may impede their ability to choose or even move away from a target.  We are beginning to distance our emotions from our business of war which could certainly be beneficial  to the effectiveness of our warriors. Yet it may also lead to very obscure issues within the wartime effort. War crimes will likely change and distribution of blame will likely be handed out in cases where a robot breaks a rule of war, or a glitch happens that causes the robot to act inappropriately. It is likely, however, that these will not happen as much as they do in cases of human mistakes due to the robots’ better vision and quicker, more precise decision making abilities.

The future of war that Singer posits in his discussion, his book, and his journal articles all seem to imply a great change in how war is fought. Yet the most important notion that he provides is likely in the form of the changing war crimes and how society will be affected by these changes. It is difficult to hold a robot responsible when a glitch happens, and our society is going to have to be prepared to place the blame elsewhere. Ultimately the change that has begun to take shape in our most recent war efforts will continue to grow and expand; it is no longer a piece of science-fiction but a reality much like the atomic bomb was, and therefore we must be better equipped to deal with the coming implications of war advancements that we have in the past.

 

Works Cited

P. W. Singer, “Military Robots and the Laws of War,” The New Atlantis, Number 23, Winter 2009, pp. 25-45.

Personal Robots and the Impact on Privacy

Posted on by
Reply

As an ever increasing amount of robots appear in our personal and social lives, it leads to many questions regarding the moral, social and psychological implications that come with this new territory in the history of human existence.  This post will discuss the possible implications of the future in robotics.  More specifically, it will discuss the implications that this increase in robot technology will have on human privacy. 

Robots have generally been seen as an aid for humans to do unpleasant or difficult tasks, and are at a rapid rate becoming used in the homes of citizens more and more.  As of the year 2007, there were 4.1 million robots working in individual citizen’s homes alone (Calo, 187).  This large increase in the use of personal robots that replace many tasks that were previously carried out by humans, leads to the possibility of a lack of privacy to citizens due to surveillance and recording devices commonly installed in robots.  This is explained clearly by M. Ryan Calo in the article Robots and Privacy; “It is not hard to imagine why robots raise privacy concerns. Practically by definition, robots are equipped with the ability to sense, process, and record the world around them” (187).  The idea of a ‘historically private home’ may be a thing of the past according to some academics as personal robots increase.  The psychological harms associated with the increase in personal robots stems from the notion that humans require alone time and due to the fact that people often anthropomorphise robots, it leaves them little to no time to get away from daily emotional stimulation when they are in the presence of others.  This absence of alone time could cause conformity, discomfort and in general psychological harm (Calo, 196). 

Socially, personal robots and the moral implications that come along with them such as the lack of privacy to citizens is still being debated in society.  Currently in Canada a similar debate that is relevant to this discussion of personal robots is being deliberated by government through the introduction of Bill C-30 by the Conservative government.  Introduced as “The Protecting Children from Internet Predators Act”, this new legislation would allow police and the government to access individual’s internet usage history through ISPs (internet service providers) without a warrant.  It would also allow police to track cell phone usage and histories, as well as installing video surveillance into the devices (Babad, 2012).  Also dubbed as the ‘snoop and spy bill’, many people argue that this invasion of privacy shouldn’t worry citizens because you have nothing to fear if you have nothing to hide (El Akkad, 2012).  This idea spouted by Conservative members of government sound eerily similar to slogans that were used to control the masses in George Orwell’s 1984 in which the figure known as ‘Big Brother’ was always watching you.  Vic Toews, Canada’s Public Safety Minister and member of the Conservative party even went so far as to state that “Canadians can either stand with the government, or with the child pornographers” (El Akkad, 2012).  This is not only an invasion of our privacy but can also be seen as an attack on those Canadian citizens who are against it since it is possible to deal with child pornographers without such extreme invasion to citizens.  This legislation caused a major uproar with the public even causing the creation of a petition titled ‘stop spying’ which is available at this site; http://stopspying.ca.   

Big Brother Canada: Harper’s Bill C30

Although these ideas of personal robots may seem unrealistic for some, it is easy to see how this is growing from the government of South Korea’s goal to have a robot in each home by the year 2015 (Calo, 196).  One example of a personal robot is WowWee Rovio which is a robot available commercially and is generally used for security or entertainment.  It is easily controlled remotely through a website which allows for both video and sound to be displayed (Calo, 192).  These home robots “in particular presents a novel opportunity for government, private litigants, and hackers to access information about the interior of a living space”, as well as personal information and information on behaviour of an individual (Calo, 188).  It is up to us to determine how far invasion of privacy will go with this increase in technology currently intertwining with moral concerns, and what sort’s boundaries need to be created to protect the individual citizens of society.       

Sources:

Babad, M. (Feb.15, 2012). “Stoop-and-scoop a better name for snoop and spy law” The Globe and Mail. < http://www.theglobeandmail.com/report-on-business/top-business- stories/stoop-and-scoop-a-better-name-for-snoop-and-spy-law/article2339143/ >

Calo, Ryan M. (2012) “Robots and Privacy”. In P. Lin, K. Abney, & G. Bekey, Robot Ethics: The Ethical and Social Implications of Robotics (pp. 187-201).  Cambridge, Mass: MIT Press.

El Akkad, O. (Feb.15, 2012). “Snoop and spy bill could be costly overreach”. The Globe and Mail. < http://www.theglobeandmail.com/news/technology/tech-news/snoop-and-spy-bill-could-be-costly-overreach/article2339111/ >

The Human Body – Perfectly Imperfect

Posted on by
Reply

The human body is simply put, the most complex and complicated structure on our planet and without a shadow of a doubt, I must begin by saying that there is plenty to be thankful for. This leads me to my main argument that we should appreciate the human body we were given and stay within our natural limits. Technology may be used to help our bodies, but only do so in ways that maintain the nature of the human body.

Simple biomechanics dictate that our body was meant to operate in certain ways, yet due to the complexity of our bodies, we have the ability to deviate from these movements at times, though these deviations are strongly advised to be few and far between. Although we are able to bend our back and hunch over, it is simply not the way that the human back was meant to operate. Although we have the ability to break down artificial and salty foods in our digestive systems, nature intended for our bodies to be digesting natural whole foods such as fruits and vegetables.

These deviations are no doubt concerning, but as technology becomes more advanced and make our lives easier, the human body will surely be having to make many unnatural adoptions. Today we sit on chairs staring at computers all day with little movement. Our fingers, once meant for grip and hold are now poking away at keyboards. Our thumbs are simply not meant to be texting.

Yet due to the ease of new technology, we do it, and we do it happily. That seems to be the new human motto these days: ‘If it makes things easier, let’s do it! Natural or not.’

Let’s take this notion a step further, humans are willing to not only adapt, but physically change their own body if it’s deemed to themselves as beneficial. People are paying thousands for laser eye surgery to artificially allow people to see. People are using advanced artificial proteins to increase body size and even attempt to change body type. Birth control pills are another obvious example of artificially changing the body for perceived benefit.

Now we reach the critical period, the tipping point, and the very place where we are today. Technology will no doubt force our bodies to adapt and our bodies will no doubt be changing as technology improves, but where will we draw a line. When will humans end their pursuit of bettering themselves?

Kevin Warwick (2011) discusses the concept of giving robots biological brains, which leads the next natural step of humans taking robot elements into their brains. Will we install chips in our brain so that we can ‘google’ things in our minds?  How long will it be before hard drives replace our memory and before encyclopaedias replace our knowledge? As shown by the turing test, artificial intelligence is only getting better at understanding the human brain and it is only a matter of time before it delves itself into human biology.

If you don’t believe me, here’s some video proof: watch?v=vb3cPjkjbi4

Now I am fully aware of how much of a jump this is and how ridiculous it seems at the current moment and critical readers should be having these thoughts while reading this. Yet, to be fair, imagine what would someone say 30 years ago, if a parent mentioned to their kid getting their first pair of glasses – ‘Don’t worry dear, by the time you’re 40, doctors will be able to shine lasers in your eye and within days you’ll have perfect vision for life!’

The point is, imagination is slowly becoming reality and basic human greed and curiosity is leading us to explore almost every possibility. This is why we must decide now where it is we will stop. Will we allow computers for minds? Will we get to pick and choose every detail and feature of our children? Will we evolve into super humans?

I for one believe that we are blessed to have the bodies we were given, and when it comes to improving ourselves within natural limits, I’m all for it. However, we should not be taking what we have for granted and we must not be greedy in our lust for self improvement. We as a race have been living on this earth for tens of thousands of years and we seem to be doing just fine naturally. Why mess with success?

What we must be cognisant of is that if all that we have serves purpose, then perhaps all we don’t have serves purpose as well.

Therefore, the body truly may be perfectly imperfect.

 

References:

Sparrow, R. (2012). Can Machines Be People? Reflections on the Turing Test. In P. Lin, K. Abney, & G. Bekey, Robot Ethics: The Ethical and Social Implications of Robotics (pp. 317-332). Cambridge, Mass: MIT Press.

Warwick, K. (2011). “Robots with Biological Brains.” In P. Lin, K. Abney, & G. Bekey, Robot Ethics: The Ethical and Social Implications of Robotics (pp. 317-332). Cambridge, Mass: MIT Press.

 

The Uncanny Valley – From Familiarity to Strangeness

Posted on by
Reply
The Uncanny Valley

A simplified version of the uncanny valley

The uncanny valley is a hypothesis in the field of robotics which states that when humans replicas act and look almost, but not perfectly, like actual human beings it causes a revulsion among human observers. In his article “The Uncanny Valley”, Dr. Masahiro Mori explains the reasoning behind his hypothesis and attempts to explain why humans feeling a sense of familiarity, or a sense of strangeness towards robots. Our sense of familiarity is easily taken over by strangeness. Familiarity and strangeness can be attained by various methods, whether it be by design, the way it acts, and our perception. Different mixtures of features can trigger a sense of familiarity, and strangeness which is when the robot dips into the “uncanny valley”.

The Uncanny Valley :

Humans have a need of understanding their surroundings, and more importantly we want to feel a sense of familiarity. To better understand this, we can use the example of climbing a mountain. It’s a function that does not increase continuously. As robots appear more humanlike, our sense of familiarity increases until we come to a valley (Mori, 33-35).

A very human like prosthetic hand

A very human like prosthetic hand

Since humans have reached an unprecedented level in robotics, we dedicate a lot of effort into creating humanoids. For example, we can easily paint over a robot’s arm that’s made of a metal cylinder and bolts to make it look human like. According to Dr. Mori, this causes an increase in familiarity. This argument states that we can achieve familiarity by making them look human, but what happens when they don’t “feel” human? Prosthetic hands are a good example. They might look real, which evokes a sense of familiarity, but when we shake the hand we know that it’s not real. In this case, we have a sense of strangeness. “In mathematical terms, strangeness can be represented by negative familiarity, so the prosthetic hand is at the bottom of the valley. So in this case, the appearance is quite human like, but the familiarity is negative. This is the uncanny valley.” (Mori, 33-35) Therefore, familiarity can easily be taken over by strangeness when we are deceived by “almost perfect human like features” that aren’t matched with the “feeling” we’re used to associating it with.

 

Positive or negative sense of familiarity?

Definitely creepy

Definitely creepy

To avoid the feeling of strangeness, we must greatly consider the design of robotics. According to Dr. Mori, we should use the first peak as the goal in building robots rather than the second. This is important because if we are to build robots that are non human like, we can more easily achieve a sense of familiarity. In the article he gives suggestions to future designers. He suggest using glasses, and maybe wooden hands could serve as a reference for future design because there is no sense of uncanny when looking at a Buddha statue (Mori 33-35). Since these aren’t human features, a sense of familiarity would be more easily attained if we gave them human like movements.

This design definitely ignites a feeling of strangeness :

To conclude, the feeling of familiarity is easily overpowered by the feeling of strangeness, and when creating robots, we must consider these factors. Robots are easily considered to dip in the “uncanny valley”. This can be caused by their exterior design, the way they act, our perception (strangeness or familiarity), and their movements. We humans have long deemed ourselves as the dominant species, and we have this sense of superiority on earth. As soon as we create robots that resembles the way we act and look, the feeling of “strangeness” can maybe be explained by a sense of threat and lack of knowing/understanding, and that Dr. Mori is right when he claims it might be important for our self preservation.

 

References

Mori, M. “The Uncanny Valley.” Energy 7:4, 1970. 33 – 35.

 

 

 

 

 

 

 

Droning On

Posted on by
Reply

Last week we discussed the arrival and meaning of unmanned technologies in warfare. Even in our classroom opinions were divided on the usefulness of such technologies, especially during the discussion on Predator drones.

The principle argument for the use of the drones is that, because they are not flown by human pilots, and because they remove the need for as many surveillance teams on the ground, drones will save the lives of soldiers by doing the same jobs with no risk of allied human deaths.

The main argument against the use of drones is that, because the operators who make the final kill decisions are miles away, far from the theater of battle, that the brutal and horrific psychological effects that come from killing another human being are removed—and war becomes a detached sort of video game.

The argument for deploying drones is much harder to defeat and that is probably the reason that drones are in use by the U.S. army today. In a time where every American soldier’s death on foreign soil seems to sway public favour away from the wars, it makes perfect sense from a military perspective to take the front-line grunts out of the equation and leave the killing to the robots.

The argument against is not even a really convincing argument, as there is plenty of evidence to suggest that the operators sending the kill orders to the drone are indeed negatively psychologically affected by their choices and they suffer as a regular soldier would.

Yet, at the end of the class I still felt that predator drones are a negative advance in the arena of warfare. This could be because of the cowardly aspect of the drones—the idea of fighting an enemy from whole continents away seems very dishonourable and frankly, unfair.

But even as I thought about this, I rejected it as myself becoming overly archaic, coming into contact with a form of warfare that was not around when I was younger. Drones are simply the result of a change in warfare that warriors are unused to engaging. The ancient Greeks believed archers were cowardly because they could fight from a distance. I am sure the Japanese thought the use of the atomic bombs in the Second World War was a dishonourable tactic as well. This is exactly the same thing as drones; the only difference this time is that we are expanding the distance from which we can strike to even greater distances.

In the essay, “Killing Made Easy,” from the textbook, Noel Sharkey argues that most soldiers are hesitant to kill another human, saying:

“Army historian Brigadier General Marshall ([1947] 2000), following after-action interviews with soldiers in the Pacific and European theaters of operation during World War II, claimed that only about 15 to 20 percent of riflemen were either able to or willing to fire. This means that around 80 percent of the U.S. infantry in World War II either were not firing their weapons when they could see their enemy, or were firing over enemy soldier’s heads. There have been some very sharp criticisms of Marshall’s research methods, and the exact percentages may not be correct, but the nature of his findings—that many soldiers are unwilling to kill—has received general support from other analyses of historical battles.” (Sharkey, 111)

This further reinforced my belief that it is not the distance of the drone operators that makes drone warfare unsettling.

It seems that the true reason that drones appear to be especially unethical is the inability of the drones to distinguish satisfactorily, due to a lack of technology and the nature of the guerilla warfare the drone’s enemies are practising, to distinguish a combatant from a civilian. Pakistani news reports estimate that “drone attacks have killed fourteen al-Qaeda leaders, and this may have been at the cost of over six hundred civilians.” (Sharkey, 115) The television episode we watched in class showed how this kind of civilian killing is both easy and simultaneously the best argument against drones, as it is is difficult to attribute blame to the killer.

Though drone operators may still be wary of killing, and though drones are undoubtedly saving the lives of American soldiers, it is really the fact that drone operators are too far removed from the subjective and unique situations of wartime encounters to act in an ethical way that is the reason they are unnerving.

But, of course, all war is unethical, so it becomes difficult to say that one aspect of the war is immoral over another, when the entire conflict is just as derelict.

 

Works Cited:

Sharkey, Noel. “Killing Made Easy.” Robot Ethics: The Ethical and Social Implications of Robotics. Lin, Patrick, Keith Abney, and George A. Bekey, eds. Cambridge: The MIT Press, 2012. 111-115. Print.

The Utility of the Turing Triage Test

Posted on by
Reply

In his article “Can Machines Be People? Reflections on the Turing Test”, Dr. Rob Sparrow argues for a particularly strong test using which, he claims, one can determine whether a machine ought to be granted moral status. The thought experiment describes a hospital’s Senior Medical Officer who, faced with a power shortage in the intensive care ward, must make a triage decision in two separate situations. The decision-making processes of the hospital are aided by an AI that “can pass the Turing Test with flying colours” (Sparrow 3). The situations are described largely as follows:

Situation1: The Senior Medical Officer must choose between two patients, the hospital’s power supply being incapable of supporting both.  Neither patient can survive without the life support system, so the Senior Medical Officer’s decision will inevitably result in one of their deaths.

Situation2: The Senior Medical Officer must choose between one patient and the hospital’s AI system, the hospital’s power supply being incapable of supporting both.  Both the patient and the AI require the hospital’s dwindling power supply to survive, so the Senior Medical Officer’s decision will inevitably result in one of their deaths.

The machine will be considered as having passed this so-called “Turing Triage Test” if and only if Situation1 is of the same character as Situation2. That is, if the gravity of the moral dilemma is equivalent in the two situations, the machine is considered as having passed and, Sparrow argues, can reasonably be considered a person.

[image credit: http://upload.wikimedia.org/wikipedia/commons/thumb/f/f6/HAL9000.svg/256px-HAL9000.svg.png]

In addition to giving sound medical advice, the AI would be capable of passing the Turing Test and, hopefully, giving its rendition of “Daisy Bell”.

Gauging the degree of moral equivalency between Situation1 and Situation2 is not an easy matter, and if passing the test is to be, as Sparrow describes it, a “necessary and sufficient condition for granting moral standing to artificial machines” (Sparrow 8 ) it must propose some kind of standard that can be used for this purpose.  Sparrow anticipates this difficulty in his discussion of the quality of the respondent.  If the competency of the respondent is not taken into consideration, then the test falls into relativism, as every individual will have a slightly different analysis of the test – resulting in the uncomfortable consequence of AI Rights being contingent upon the mere opinions/intuitions of policymakers.

To respond to this challenge, Sparrow makes it clear that:

(1)   the decision-maker must be a rationally competent individual that, given the gravity of the moral dilemma, could understandably find themselves feeling regret or remorse toward their action (which protects against objections appealing to sociopaths, etc.).

And,

(2)   The quality of the respondent’s intuition must be independently considered (which ensures that the responses to the test are sound, not relying on mere intuitions that may or may not be in accordance with morality).

[Image credit: http://www.cmu.edu/news/image-archive/keepon_236x236.jpg]

Guided by moral intuition alone, the danger is that someone might be willing to ascribe higher moral status to a robot like Keepon than a vastly more intelligent AI like IBM's Watson, merely due to their bias toward cuteness.

Although this final stipulation protects the Turing Triage Test from falling into relativism, it comes with significant cost. Conceivably, the tools and arguments developed for evaluating the quality of respondents’ intuitions could be used for answering the question itself.  Consider the two following objects of analysis:

(1)   Whether respondent X’s answer to the Turing Triage Test was of sufficient quality

And,

(2)   What the highest quality response to the Turing Triage Test would be.

It seems plausible that the tools that Sparrow admits would be necessary for approaching (1) would be easily transferrable to answering (2) instead.  The test would then be of value for use as a foil against which these tools may be developed, but insufficient as an approach to studying the problem.

Video: Philosopher of mind Daniel Dennett on the Turing Test which Sparrow’s Turing Triage Test is supposed to be an adequate replacement.

To conclude, I appreciate what Sparrow is trying to do with the Turing Triage Test when it is used as a tool for communicating the issues of AI Personhood in a way that is both accessible and clear to the interlocutors.  Further, I am inclined to agree that, given his stipulations, the test can be seen as expressing necessary and sufficient conditions for the extension of personhood rights to AI systems.  The arguments surrounding AI Personhood are, however, focused on the latter of his stipulations – the quality of respondents’ intuitions – and it is in this area that further work is truly needed. The tools and arguments used to evaluate the quality of these intuitions can be directed toward analysing whether Situation1 and Situation2 are morally equivalent. If it is determined that they are equivalent, then the test portrays an interesting consequence: that killing a machine(person) would then be equivalent to killing a human(person).

Works Cited

Sparrow, R. (2012). Can Machines Be People? Reflections on the Turing Test. In P. Lin, K. Abney, & G. Bekey, Robot Ethics: The Ethical and Social Implications of Robotics (pp. 301-315). Cambridge, Mass: MIT Press.

Vote Yes for Robotic Servants!

Posted on by
Reply

The concept of ‘robotic slaves’ or servants are appealing to many

The perfect robot servant, minus the attitude

because of the considerably high convenience it offers. Having robots to clean the house, make dinner, and perhaps even visit your significant other parents for you sounds ideal.  The robots I am referring to are the ones that would resemble people, in that they can think and engage with the environment by making their own decisions. One possible ethical dilemma that arises however is the concern of giving all the ‘undesirable’ jobs to robots, while humans reap the benefits of success. I believe that Steve Petersen effectively countered this problem in his piece Designing People to Serve. In this writing Petersen goes on to outline the three criterion for avoiding ethical problems with robots doing all the ‘crappy jobs’. They are…

1)      The robots are people in every relevant sense

2)      They comply with our intentions for them to be our dedicated servants

3)      They are not being wronged during the task

Robotic servitude may sound unpleasant to some because of the negative connotation attached to “servicing” someone. In other words, being someone’s servant usually involves menial tasks that are often repetitive and dull. Robots could be ‘people’ in every sense such as enjoying the various rights that humans do; however the cognition both experience could be vastly different.  Petersen goes on to stress this point when he writes that

“… we humans get our reward rush of neurotransmitters from consuming a fine meal, or consummating a fine romance… If we are clever we could design [artificial people] to get their comparable reward rush from the look and smell of freshly cleaned and folded laundry…”         (p.284). If a robot gets enjoyment out of cleaning the sewers or cleaning toilets, it would be more of an ethical dilemma to prevent them from completely such a harmless task. Just because a task is unappealing for humans and does not offer a significant satisfaction on completing it, doesn’t mean that artificial people will feel the same way.

Maybe a bit overboard, but the idea of robot servants is clear

In finding meaning in life, I believe that artificial people would be considered immanentists. In other words, meaning would be derived from within life itself. This is important because if robots could fulfill their meaning, which could be as simple as cleaning dishes all day, then we shouldn’t frown on allowing them to accomplish it; of course programming the robots to feel a type of pleasure when cleaning dishes doesn’t hurt in guiding an artificial person in finding their meaning. This is where I make my distinction from robots being servants to robots being slaves. As Petersen mentioned before, we could design artificial people to find reward in the tasks that are unsafe or inconvenient for humans but we don’t have to force them to do it; that’s not to say that a sense of achievement or fulfillment couldn’t be found in other activities.

Peterson goes on to describe the different levels of pleasure and how artificial people, just like humans, could indulge in them. After an artificial person “… [binges] on their baser desires [like] washing laundry, the sated Aps will then turn to Shakespeare or Mahler for a while.”(p.293) This quote from Petersen is vital in understanding that if we create an artificial person that desires to fold laundry, that were not just creating a slave purely for our convenience. We are creating a person that has its own desires and dreams, yet still has an intrinsic need to complete tasks that benefit humanity; even if it’s just as something simple like shovelling the snow. In closing, robotic technologies should be pursued so that they both benefit the robot, by giving it qualities that humans possess, but also humans, by making everyday life more convenient.

http://www.youtube.com/watch?v=nXKGlJIL7-o

- Trailer for a robot documentary that looks pretty informative

http://www.youtube.com/watch?v=S31X2fSpbfI

- Skip the first two-thirds of the video to where Cleaning robots that were given faces; still no where close to a person, but the idea of giving a cleaning robot a face is interesting.

 

Petersen, Steve. (2010). “Designing People to Serve.”. In P. Lin, K. Abney & G. Bekey (Eds.), Robot Ethics (pp. 283-298). Cambridge, Mass: The MIT Press.

Big Brother

Posted on by
Reply

Let’s imagine a future for a moment, one where every household has a robot. It helps with dull household tasks, such as the laundry and cleaning, generally making human lives easier. An item that was unthinkable years ago has become so commonplace that no one can imagine a life without one. In fact, people are beginning to forget how to do many of the simple tasks that their domestic robot has taken over. The way people think and interact with the world has changed, not unlike what has happened with computers and other similar devices in the past.

Any individual connected to a computer or in possession of a smart phone knows that their life is on display, to fellow individuals, corporations, and the government – often without their express knowledge. We currently discuss the implications of government access to things such as an individual’s computer or ISP address, as we see with Bill C-30 in Canada and in anti-terror legislation in the United States. Someday if robots become a common part of the domestic landscape, a similar discussion will resurface. Whatever balance is struck in the debate surrounding computers and the internet in regard to privacy will most likely have a large influence on similar legislation regarding robots. If this is the case, then it will be important that we strike the right balance now, as unlike computers, which may only see what the individual types in; domestic robots would see the private everyday life of individuals inside the home.

With every additional piece of technology which connects us and promises to streamline our lives, privacy may be compromised. There are an increasing number of avenues through which details of a person’s life can be illuminated, such as the internet and smart phones, this invasion of privacy would be compounded by robots. In the chapter Robots and Privacy they state that “practically by definition, robots are equipped with the ability to sense, process, and record the world around them” (Calo, 187). Beyond robots which may be utilized by the military and law enforcement for surveillance purposes, domestic robots owned by private citizens offer a great risk to privacy as they offer a new point of access to historically protected spaces. As the chapter asserts, even a description of how a person programs and interacts with a robot may read like a session with a psychologist (Calo, 188). There is no doubt that it could become difficult to self-censor around a robot that is part of your everyday home life. This could be pretty heavy stuff, especially when combined with all the other ways information can be collected about people presently. While I don’t believe that we should resist progress, I do believe that we should work toward striking a fair compromise between the rights of the government to access information and the rights of citizens to privacy.

The balance seems to have tipped away from citizen’s right to privacy toward the government’s right to watch, to use the United States as an example, evident in the USA Patriot and Homeland Security Acts (Weinstein). While these acts may have the intention of fighting against terrorism, they have wide ranging effects on privacy for all citizens (Weinstein). The debate continues however about correcting that imbalance and determining what the rights of citizens and governments should be. It is the outcome of that debate which may influence the future.

In conclusion, if robots become common in the home there will, in all probability, be no area of an individual’s life which cannot be monitored by “big brother”. While it may sound slightly paranoid, the current debates regarding privacy on the internet demonstrates that maintaining a right to privacy for citizens is becoming a serious concern. The outcome of the debates about a fair compromise between citizen rights to privacy and a government’s rights to watch will in all likelihood have a significant impact on the future rights in regard to robots. It is the case then that all citizens should pay attention to the laws governing privacy which exist or may be adopted in the future to ensure that proper balance is struck  to continue to ensure the privacy of individuals, especially as an increasing number of parts of citizens lives are on display.

Works Cited

Calo, R. (2012). Robots and privacy. In P. Lin, K. Abney & G. Bekey (Eds.), Robot Ethics (pp. 187-201). Cambridge: Massachusetts Institute of Technology.

Weinstein, L. (2002, December 30). Year in privacy:  Citizens lose. Wired, Retrieved from http://www.wired.com/politics/security/news/2002/12/56954

Moral Risk and the Turning Triage Test

Posted on by
Reply

A surgical robot.

1. The Turing Triage Test [1] is designed to help us determine whether or not a machine has moral standing. According to the test, if we experience a true dilemma about whether or not to sacrifice a machine to save a human life, then the machine has achieved moral standing and must be treated as such. However, I will disagree with this conclusion. The test does not tell us whether or not the machine has moral standing; yet, it does tell us whether or not it must be treated as though it does.

2. The test does not concern itself with whether we choose the human or the machine, only that we experience a difficulty in making that choice. Sparrow also suggests that we must assume the person making the choice is reasonable. An unreasonable person might be biased toward or against the machine and not experience the dilemma, or might experience the dilemma even if the machine were actually a toaster, damning their credibility. This is an important point because the way we reason about personhood is applicable to both humans and machines.

3. To avoid metaethical fog [2], the test assumes that personhood is sufficient for moral standing. If this is so, then the question of whether or not the machine has moral standing is really a question about its personhood [3]. Consciousness seems to be an appropriate prerequisite for personhood. Unfortunately, we have no way to access the machine’s mind, if it exists, in the same way we access our own, and thus cannot determine whether or not the machine is conscious. This is true for the human too. We know only of each other’s consciousness through comparisons between how we behave when we think certain things and how others behave. That the other is conscious is induced from these experiences.

4. If we assume that another human is conscious based on its behaviour, then we must assume that a machine is conscious if its behaviour is similar enough. Further, because we assume humans are cognitively developed enough to have moral standing (and moral standing is a consequence of cognitive ability, rather than, say, God), then any machine which appears to be as cognitively developed, or nearly as developed, as a human must also be assumed to have moral standing.

In, “I, Robot,” Detective Spooner is conflicted about Sonny’s personhood.

5. To not treat the machine as though it has moral standing, even though it acts in a manner similar enough to the manner we deem sufficient for humans to have moral standing, is to be contradictory. Consider, for example, an alien species which visits Earth. They have very different physiologies than us, but appear intelligent. We would almost certainly ascribe moral standing to them. The machine has a different physiology than us, but if it appears to be able to cogitate at a relatively human level then we must ascribe it moral standing just the same.

6. This does not mean the machine has moral standing. It only means that we must treat it as though it does. The test does not determine whether or not a machine has moral standing. How we treat others, or think about others, is not sufficient to make them moral agents. To be persons they must be conscious, but we cannot will something to be conscious, nor does the way we think about it change the fact that it is or is not so.

The medical hologram aboard Voyager is often treated as a member of the crew.

7. Even though we must act as though the machine has moral standing, there may still be circumstances in which it is preferable to sacrifice the machine. A machine may act similar to a human, but not exactly the same. This brings us to doubt whether or not it is conscious, thus we doubt its personhood, to a greater degree than we would for another human. We must still treat it as though it has moral standing, though. The chance that we act immorally in a situation I will call ‘moral risk (related).’

The robot on a moonbase in the movie, "Moon." Unfortunately, due to spoilers, you'll have to watch the movie to determine the relevance!

8. In the first round of the test, there is very little moral risk as most people do not doubt that humans are conscious beings with personhood. In the second round, however, there may be greater moral risk when we consider the machine. We doubt the machine’s personhood much more than a human’s, thus there is a greater chance we act immorally if we sacrifice a human to save the machine. The moral risk of choosing the machine over the human is higher. If the machine is sufficiently human-like, then the moral risk is much lower. We must also consider the human’s chances of survival versus the machine’s.

Sources

[1] Sparrow, Rob. Can Machines Be People? Reflections on the Turing Triage Test. Robot Ethics: The Ethical and Social Implications of Robotics. Edited by Patrick Lin, Keith Abney, and George Bekey. Cambridge, Massachusetts, USA: The MIT Press, 2012.

[2] Abney, Keith. Robotics, Ethical Theory, and Metaethics: A Guide for the Perplexed. Edited by Patrick Lin, Keith Abney, and George Bekey. Cambridge, Massachusetts, USA: The MIT Press, 2012.

[3] Asaro, Peter. A Body to Kick, but Still No Soul to Damn: Legal Perspectives on Robotics. Edited by Patrick Lin, Keith Abney, and George Bekey. Cambridge, Massachusetts, USA: The MIT Press, 2012.

Human Morality and Decision Making

Posted on by
Reply

http://perceptualunderstanding.com/tag/consciousness/

This post considers the implications of duplicating human morality towards decision making. This is relevant to roboethics because replicating human morality offers some insight towards programming a consciousness for a robot or software. A person’s past experiences and knowledge shapes their morals and values. Ethically speaking, morals would depend on what is right or wrong and it would offer stability in terms of guidelines for deciphering what is ethically correct or immoral. In order to have robots making morally sound decisions, human morals need to be analyzed. A person must be self-aware of the current situation and the surroundings in order to fully comprehend a correct pattern of decision making in their consciousness. There is an essence of perfection in the midst of accepting the right decision path for a practical, but moral outcome. A robot can not make a fatal mistake in their pattern of decision making, therefore the programming must be accurate and be as close to perfection as possible.

http://www.dailygalaxy.com/my_weblog/2009/07/future-of-robot-intelligence-could-ant-societies-be-the-model.html

In the Allen and Wallach reading about moral machines, they support the idea of replicating an aspect of human morality when they state “it is necessary […] to make them capable of factoring ethical and moral considerations into their decision making” (2012, p. 56). They further explain that the need of moral standards will depend greatly on how much freedom a machine will have when it comes to creating a consciousness. There has to be a sense of growth and expansion in terms of freedom and how one acquires it so that a machine does not appear to be out of control if morals no longer apply. An example could be the scene from the movie A.I. Artificial Intelligence where the character David loses his temper and does not consider any humanistic morals that he might have in place (“A.I. David’s breakdown”, 2012).

Another aspect in the reading that they discuss is the self-organizing phenomena of morals as they state that “cooperation and a shared set of moral instincts might emerge” (p. 59). Moral instincts can come from past experiences and current surroundings that can affect future decisions (Ashman, Carver, Dawson, 2000, p. 697). A bad experience can alter the outcome of a potentially good moral situation. Humans do tend in fall into patterns, but at times they can be unpredictable. If human morals were duplicated to enhance a machine’s decision making process then this is an important issue to consider. An example of a past experience affecting the present can be seen in the film Blade Runner where the character Rachel realizes that she is a replicate and that her memories were programmed (“Blade Runner: Rachel realizes her memories aren’t hers”, 2012). This clip demonstrates the capability to contemplate reality. This character is deciding whether her identity is an extension of reality or just programmed. She could potentially organize new morals based on this new information.

Perfection is important to enforce the ideals of the right decision in terms of morals and values. In the Hughes reading about a Buddhist approach, he speaks of moral behavior when he states that “all moral behavior flows from the struggle to perfect them” (2012, p. 77). Accuracy is important to infiltrate a perfect moral compass in a machine’s consciousness so that a decision appears to be correct for the situation. One also needs to be self-aware of being present in the situation in order to come closer to perfecting the correct decision making process for a machine. An example of perfection can be seen in the clip where a robot recognizes its current situation and its surroundings in order to complete the task of pouring a glass of water (“ Robot That Can Learn, Think and Act By Itself #DigInfo”, 2011).This demonstrates the human means of striving towards perfection and at the same time perfecting the thinking process of the robot.

All in all, it is evident that duplicating human morality towards creating a consciousness for a machine depends on the aspects acquired throughout the decision making process. This is seen through past experiences, knowledge, surroundings, self-awareness, and of course the achievement of perfection. All of these aspects contribute to creating consciousness by simulating human morality.

References

Ashman, S., Carver, L., & Dawson, G. (2000). The Role of Early Experience in Shaping Behavorial and Brain Development and its Implications for Social Policy. Development and Psychopathology, 12 (2000), 695-712.

Hughes, J. (2012). Compassionate AI and Selfless Robots: A Buddhist Approach. In Robot Ethics, 69-83. Cambridge, MA: MIT Press.

Wallach, W., & Allen, C. (2012). Moral Machines: Contradiction in Terms or Abdication of Human Responsibility. In Robot Ethics, 55-68. Cambridge, MA: MIT Press

ASIMO Can Steal My Dance Moves Any Day

Posted on by
Reply

Scott Feschuk wrote a great post in Maclean’s this week. Very funny. I like that the robot warrior gets the friendliest rating of all the robots. He manages to touch (very briefly) on many of the oddities of robotics, as well as the scarier aspects. We’ll be studying these issues in more detail throughout the course.

I for one would be happy to have ASIMO steal my dance moves. If imitation is the best form of flattery, then ASIMO represents the ONLY hope that I’ll someday look good on a dance floor.

Humanoids vs Socially Interactive Robots

Posted on by
Reply

Robots the Show Emotion

With the rapid advancement in robotics, a time has come where the ethics of constructing and implementing robots within our society is a very real and legitimate concern. It is quite obvious that the advances in robotics is leading technology towards creating robots that are more human-like, as is the case with Dr. Hanson’s empathetic bot. There are many ethical questions to consider going forward in robotics, such as the social cohesion between robots and humans, yet the one I want to focus on is more of an immediate issue. I will look to focus on the ethical concerns of creating robots that empathize, or if continued on this path potential sympathize, and create their own emotions.

The robot shown in the video above by Dr. Hanson is hardly able to comprehend the emotions evoked by Hanson, in the Searlean sense at least. He does however show the ability to mimic these emotions which, for now, will have to suffice. If one were to test this robot in a sort of facial-emotive Turing test, it may very well pass, but we would most certainly be wrong in saying that this robot feels these emotions with its human companion.

It is likely fair to say that most people, when feeling down, appreciate the empathy or the appearance of empathy from someone close to them. Of course, it is more likely that the person empathizing would be more effective based on the strength of their relationship with the person who is upset. Based in this notion alone the robot presented by Dr. Hanson may have a role to play in our society as far as robot-human relation at the very lowest level is concerned. The concern here is how these bots will continue to be programmed. Bekey speaks of a future when robots “develop distinctive personalities and communicate with each other.” (Abney, p.29) It seems like a large advancement but considering where technology is at already with the mimicking of emotions portrayed in Hanson’s bot, and the advancement forecasted by Moore’s law (Lin, p.3) it is certainly in the foreseeable future.  I want to argue though, that we have gone far enough.

A robot that is able to mimic a person’s emotions insofar as it makes the person feel better, or at least less alone in their pain or more excited through the joy of sharing happiness with another person, or robot is reasonable. A robot that has its own emotions, and can therefore create its own vocation seems almost diabolical. As if our planet wasn’t already crowded with emotional humans, were going to fill it up more with emotional robots? It seems like scientists are actually testing to see if robots can surpass human intelligence and ultimately take over as the dominant species on this planet. It sounds oddly like the Terminator, but considering the advancements and the directions in which robotics seem to be headed, it is hardly outlandish. I propose an alternative to this steady advancement.

In our society it seems like the highest demand for robots is within the grouping of “Humanoid” robots. These are the robots that do certain duties around our houses and at our workplace, like the Roomba, which is seemingly the most popular robot in the world based on its over 4 million sales. (Abney, p.25) It would be hard to disagree that these sorts of robots seem better suited for our society compared to robots that share our emotions or even create their own. The feeling of accepting the humanoids over the socially interactive bots may very well be completely attributed to the idea of the “uncanny valley”. (Abney, p.26) The humanoid bots seem somewhat like humans in their actions because they are doing jobs for us, yet the socially interactive robots, like the one shown by Dr. Hanson, seem almost too much like humans, as if they were tricking us into thinking they are human. The deception is quite apparent and will only continue to deepen the further this form of socially interactive robot advances and that in itself seems morally wrong.

I noted at the beginning that this is something of an immediate issue. This is because we have obviously hit the point where advancing further could lead to pushing society into a limbo point where adjusting to the very human-like robots will be difficult for a great many people. It is hardly something that we can allow based solely on the fact that it shows how far our technology has advanced. The moral implications of creating an entirely new species creepily similar to our own is wrong based on the implications it could have on our species. Creating companies capable of showing fake emotion for the sake of cheering somebody up or celebrating with someone, is enough advancement for robotics, and therefore all other advancements should have aims of improving aspects of our lives objectively, not adding to an already overpopulated Earth with a potentially deceptive new species.

Bibliography

Abney, Keith, and Patrick Lin. Robot Ethics: The Ethical and Social Implications of Robotics. 1. 1. Cambridge, Mass: MIT Press, 2012. 1-34. Print.

 

Robots are not people, but we can still be kind

Posted on by
Reply

 VS

In Rob Sparrow’s article he proposes a test for granting personhood to robots called the “Turing Triage Test”. It goes something like this. Imagine you are a senior medical officer who has worked closely with a robot who has proven itself a huge asset for diagnosing patients as it is able to learn, reason and make decisions. One day there is a disaster and the power is compromised. There are two patients on life support and there is only enough power to save one patient and you must choose who will live and who will die. The robot helps you choose who will be the most likely to survive and you make your decision. In a second similar scenario, there is a patient dependent on life support and the robot, who you have worked side by side with, who is also dependent on the power to continue existing. Sparrow says that only when these two scenarios have the same moral character, meaning there is the same type of moral dilemma with choosing between a robot and human as there is between two humans, will robots have moral personhood.[i] With the “Turing Triage Test” it seems almost impossible to imagine picking a robot over a human. Our intuitions may tell us that robots will not be equivalent to people on a moral level, however this does not mean that we have free reign over the way that we can treat them.

In the video Measure of a Man Commander Bruce Maddox has no trouble at all explaining sentience as “intelligence, self-awareness, and consciousness”. The danger of Maddox’s claim is that he is trying to limit a very complex question about ‘what is a person’ into three simple traits. No matter how we try to define moral personhood there will always be people who we do give moral status to that don’t fall within our definition. Small children, people with disabilities, the elderly, people in comas and even people who are sleeping could not be considered sentient persons under the definition given by Maddox. Denying the personhood of these people seems ridiculous therefore Maddox’s definition must be lacking. The beauty of the ‘Turing Triage Test’ is that it does not get caught up in defining personhood but simply points to the fact that there is something that we value about human life over robots, regardless of intelligence or capability. To value human life over robots is not a weakness that points to our cruelty towards things that are different from us, but rather an indication that there is something inherently valuable about human life.

I realize there is also a danger in the idea I have proposed. It seems as though proclaiming superiority only facilitates disrespect and even abuse of inferior things. The best example is our abhorrent treatment of factory farm animals.  In order to model what I believe would be the best approach towards treating robots with respect I will turn to the Catechism of the Catholic Church and its teaching on animal treatment.  Describing how we are to treat animals it says, “It is contrary to human dignity to cause animals to suffer or die needlessly. It is likewise unworthy to spend money on them that should as a priority go to the relief of human misery. One can love animals; one should not direct to them the affection due only to persons.”[ii] According to this quote we have a duty to treat animals with respect and make sure they are protected from suffering as long as it does not conflict with ensuring that ending human misery is the priority. It is this same attitude that I propose we adopt towards robots. This I believe is the middle ground between making robots equivalent to humans and treating them like they are insignificant. This balance will ensure respect for human life without compromising the needs of robots.

[i] Sparrow, R. (2011). “Can Machines be People? Reflections on the Turing Test.” Section 7 of Robot Ethics. 302-3

[ii] Catechism of the Catholic Church. Respect For the Integrity of Creation. S. 2418 http://www.scborromeo.org/ccc/p3s2c2a7.htm#2418

Suicidal Robots?: Asimov, Labour and Autonomy

Posted on by
Reply

"I didn't ask to be made: no one consulted me or considered my feelings in the matter. I don't think it even occurred to them that I might have feelings"

In roboethics there are a number of problems to be addressed, particularly when it comes to the ethical status of robots and their autonomy. There are considerations to be made about programming, behaviour, appearance, moral status, emotion and others which all have a bearing on the way we view and treat robots and, also importantly, the way we ought to design and program them.

Robots are currently used in a number of fields, often dangerous, messy, boring or menial jobs that we do not wish to do ourselves; jobs of this nature tend to have strong ties to poor working conditions, depression and suicide among employees. Foxconn, the main manufacturer and supplier of Apple products, for instance, has recently voiced an intention to replace a good deal of their staff with robots after a rash of employees committed suicide.[1]

Current legislation regarding robots is thin but public policies are emerging in the world, many finding their foundations in Isaac Asimov’s 3 laws of robotics. The second law states: “A robot must obey orders given to it by human beings except where such orders conflict with the first law” and the third holds that ”A robot must protect its own existence as long as such protection does not conflict with the first or second law.”[2]

The first law concerns the protection of humans as paramount and is the most agreeable in my opinion. The second and third, however, seem to expose an enormous problem for the autonomy of robots. This sort of constriction may hold negative ethical implications no matter how attractive it may seem to remain in complete control and protect your investment from itself.

So, what happens when a robot crosses or (perhaps more interestingly) straddles the line between being a tool and an actual autonomous being?

Issues like empathy and emotion come into these ethical debates along with rights, rational capacity and autonomy which lead me to consider the darker side of emotion and the existential condition of life. What if there were genuinely depressed or suicidal robots that would have no options? The first example I thought of which depicts such a machine is Douglas Adams’ Hitchhikers Guide to the Galaxy series; the character Marvin is a  robot designed to have a “Genuine people personality.” Marvin is depressed because he is never able to use the full scale of his massive intellect in the menial, routine jobs he’s given.

[youtube http://www.youtube.com/watch?v=dxp1AqWleT4&w=420&h=315]

Or, closer to the heart of the matter, consider this 2007 Superbowl ad made by GM featuring one of its factory robots becoming despondent after being fired from his assembly position. The robot tries a number of  jobs afterwards and eventually throws itself off a bridge.

[youtube http://www.youtube.com/watch?v=B3NGN4t4hm4&w=420&h=315]

The dream sequence in the commercial is obviously exaggerated and intended to make people empathize with the robot despite its relatively simple design and function in reality. Still, it isn’t difficult to make connections to poor Marvin or the workers. There may come a day when the self-awareness of robots is clearer and their moral status will be cemented alongside our own.

It is my contention, then, that it is paramount to protect whatever autonomy robots may posses in coming years. Marvin ought to have the rights to act as he wishes rather than be ordered into labour for us without recourse of action.  Programming robots to be deliberately happy seems like coercion of a similar nature; like our intuitions show in Robert Nozick’s experience machine concept,[3] there is something about reality we value and robots ought to be given the autonomy to be similarly free in this “real” way.

Robots have the seemingly infinite potential to become unique and distinct conscious beings (perhaps even with intellects vastly superior to our own, like Marvin’s) and we ought to be primarily concerned with maximizing their freedom and autonomy of action in order to provide the opportunity for such progressions to be made. Perhaps a device like a Roomba need not be respected as a moral agent (for now [4] [5]) but there is a wider scope here concerning autonomy and rights that needs to be considered in a broader light than those considerations raised by Asimov.

[2] Asimov, Isaac. Runaround, 1941

[3] Nozick, Robert (1974). Anarchy, state, and utopia. New York: Basic Books. p. 42-45.

[4] http://blogs.psychsterdata.com/yjgm/2010/03/of-robot-maids-and-human-nature-what-our-roombas-say-about-technology.html

[5] http://spectrum.ieee.org/automaton/robotics/artificial-intelligence/how-to-give-robot-vacuums-a-personality

The Idea of Understanding

Posted on by
Reply

Hello Internet World!

 

And welcome to the start of my blogging career. For this blog I really want to discuss the idea of understanding. I find that this is a fundamental issue made in both the Searle article and the Turing article(a good summary of both their works can be found here in comic beauty:http://www.youtube.com/watch?v=TryOC83PH1g) . For my blog, I really want to discuss the fundamental roots of understanding. The Turing test is designed to show understanding in regards to robots, but the effort is successfully debunked by Searle’s Chinese room. Searle argues that there really cannot be a strong artificial intelligence, because of the proof given in his article. Robots operate on codes and programming. Robots cannot perform tasks or make links to things that they were not programmed to make. I would like to argue that this is the crux of understanding from a humanistic point of view. The human mind is an incredibly subjective entity. It learns and adapts to situations in ways that are still difficult to comprehend. A perfect example of this is in an offshoot of the Chinese room test. Imagine Searle was in China, and was given all the tools previously mentioned (the codebook, the incomprehensible writing etc). The only difference to this test would be that Searle would have interactions with the Chinese people face to face. The Chinese would give him codes and he would turn them into other codes and hand them to whomever he should. Searle I would argue would get more out of this interaction. He would see the facial reactions of the Chinese in reading his work, and start understanding at a base level what he wrote down. A joke would cause his superiors o laugh, and so Searle would understand that what he wrote down was funny. Seale takes with him thoughts and experiences that tell him that this is a good thing. He would then have a small base of Chinese knowledge that would systematically increase on a trial and error fashion. Granted he may not still understand exactly what is written , but he would in time come to understand the basic reactions of his observers, and perhaps even come to know the basics of sentence structure and grammar based on repetition and causal observance. Right now, the computers of our time cannot hope to get this far. They do not have the thoughts and past experiences to understand human behaviour and draw information accordingly, even with advanced sensors. I would like to argue that the act of action, observance, internal analysis based on past experiences and reaction is the basis of human learning. Computers I would argue have a simple action reaction quality to them. If A then B. Searle is actively demonstrating this in his Chinese room work. In his Chinese room thought experiment, he is the computer. He simply, has to act upon the work given to him, and reacts according to what his code tells him to do. Searle in China without the room demonstrates that there is something incredibly different about humans. So how should we react to this knowledge? Searle certainly shows that there is a long gap between having knowledge and simulating knowledge. My thought experiment should show that humans have qualities and abilities that robots do not. Now this blog will not go into the ethical debate about if we should bridge this gap. That will be left for another time. I would simply like to address the incredible difficulty in getting up to human standards. How can a robot understand a human? Sensors would help. They could and would give the computer a sense of observance to the world around them. Still, humans are irrational at times, and so there would be times where a computer would not be sure what to think. A human could cry tears of joy, of pain or of anger, and perhaps only human intuition would know for sure which was which. There would still be a problem however, with drawing from past experiences. Computers do not have that data bank. On top of that, humans react much differently to robots as opposed to humans. Uploading a bank of human history onto the circuits would therefore probably be of little use. I really do not have an answer for this one. I simply put forward in this blog that computers lack the skill of interpretation and adaptation needed to become understanding entities.

 

Until next time!

Computing Machinery And Intelligence

Posted on by
Reply

 

The paper presented by Alan Turing ,Computing machinery and intelligence presents us with a unique thought experiment meant to instil within the reader questions about what constitutes an intelligent and thinking agent. This is done by placing an interrogator between two people in which he must find out which one is human and which as not by proceeding through a series of carefully tuned questions(Turing, 1950). Turing is not making an empirical claim by any means in this experiment, he is trying to show that even if a machine is intelligent enough to fool us into thinking it is human that still doesn’t leave us with any idea what the means used to do so are, despite his speculation it might be a three part process(Turing, 1950). He is suggesting that the Turing test is the best means of determining if a machine is intelligent aside from being able to probe into its mind and analyze its software. Intelligent or not within fifty years, he is skeptical of using modern understandings of the term thinking to apply to machines in the future, there very ability to deceive us may call for a drastically new definition of the word. This is in fact one of the objections proposed, how do we know that what the machine is doing constitutes thinking, in which Turing replies we don’t know, but if it does well at the imitation game that objection is in trouble(Turing, 1950). An interesting objection in the article is the argument from consciousness which claims machine can only be considered thinking when it relates to us that it feels itself thinking. Turing responds by granting the other minds debate and says machines are no different in that we cant know there conscious anymore or less then I know my brother or my dog is conscious, the bar is to high to validate it, but the Turing test is our best hope(Turing, 1950).

 I would like to further the dialogue presented in Turing’s article by presenting alternative measures to determine the complexity and degree of intelligence in a robot, but also the extent to which a robot is conscious of its own being. The first test is actually one performed on children in an attempt to measure their own degree of self awareness, the Rouge Test. When placed in front of a mirror with a mark on their face the test proves that by the age of two kids notice the mark in relation to the face being of their own self. it(Berk,2008). It is important to note that the validity of the test is important not the method. It doesn’t seem feasible(just yet) to place a robot in front of a mirror and see if it recognizes itself, but if a test could be designed to show that a robot is capable of understanding its choices in relation to its own self  might give us some insight into its ability to think deeply and perhaps compliment the Turing test. A test like this would yield evidence parallel to humans in terms of understanding ones own unique workings. For a machine that may not be a conscious agent in a mirror but a recognition of a different nature, the knowledge that internally something has been re-routed or externally that something has broke off. This ties in nicely to Turing point that our definition of thinking may need to be drastically changed in the years to come.

The second method I would like to present is again largely theoretical but would demonstrate that a machine has an intuitive element to it. Known in philosophy of the mind, Descartes came up with the Demon Test which he said could justify the dualist perspective that mental properties exist separate from physical properties(Baker, 2011). Simply put, the thought experiment asks if a mysterious demon controlled all of your actions and bodily movements, what would be left? Descartes answer was that what couldn’t be controlled was the mental stuff, the conscious knowledge that you are aware of your own thoughts, beliefs and desires is something that the mysterious demon couldn’t control, something that you have special access to(Baker, 2011). Again such higher order thinking is most likely many years away from our current robotics but if a machine could one day pass such a test I think that we would be very reluctant to say it isn’t capable of thinking despite its method of design or its look. 

To summarize, I think that these two examples should one day be used to compliment the Turing test in determining if we are going to deem a complex robot as a non human entity or an intelligent being. It should be a knowledge of your own being and intuitive nature that leads to a status hood or some sort of ethical right. I would say that this discussion is important for robot ethics because the way in which we perceive a robots internal world is going to dictate both how we create them and also how we treat them .We are at the stage of robotics were a grey area is emerging especially in the anthropomorphic sense, and that grey area is going to get even larger if robots can soon demonstrate internally intuitive capacities no matter how genuine or artificial they may be.

 

 http://www.youtube.com/watch?v=W-pc_M2qI74

 

work cited

Turing, A.M. (1950). Computing machinery and intelligence. Mind, 59, 433-460.

Baker C, Goetz S:(2011). The soul Hypothesis- Investigations into the existence of the soul, Continuum books, pages 26-46

 Abney K,G Bekey G(2012):  Robot Ethics: The ethical and social implications of robotics, Massachusetts Institutes of technology, pages 3-13

Berk L. Infants, Children And Adolescents, 6th edition,Illinoisstate university, pages 168-173

Philosophical Zombies and Robots: What to do When Consciousness is Possible

Posted on by
Reply

In philosophy, a zombie is “exactly like us in all physical respects but without conscious experiences: by definition there is ‘nothing it is like’ to be a zombie” (Robert). Few people actually believe that the humans around us are zombies, and that makes sense, since each of us is a human being, and it seems reasonable to assert that our fellow humans, being so similar, also have conscious experiences, with all that that entails (e.g. agency). As robotic technology advances, however, a problem will emerge: we cannot tell when non-human beings have consciousness. As such, I will argue that we ought to treat sufficiently intelligent and complex robots as though they had a conscious mind. Essentially, this is because it is possible that they do, and it is better to avoid potentially harming conscious beings than it is to use them like modern computers and presume that they are not conscious.

If it isn’t already clear, this view hinges on the premise that one cannot confirm that consciousness, the internal lives of beings, exists in a machine. Solipsism, a form of philosophical skepticism, supports this, by making it clear that we cannot have certain knowledge of the external world (Encyclopædia Britannica); this includes knowing that our fellow humans have minds. There is good reason to doubt philosophical skepticism, though, since one can and should assert things that it is reasonable to assert, even though it may still be possible to doubt them (Lewis and MacDonald, 50). Still, based on this view, it is worth noting that it is possible that our fellow humans lack minds, and we have much better reason to believe that humans can have minds than machines, since the only being we can certainly claim to have consciousness, i.e. ourselves, is a human.

Turing had a similar view. Turing was so certain that consciousness cannot be determined in machines that he considered the question “can machines think?” to be “too meaningless to deserve discussion” (Turing), and it is certainly impossible to determine an answer to a meaningless question1.There have been some thinkers who have denied that zombies are possible, or even conceivable (Robert), but Searle presents good reason to deny that. Searle gives the example of a machine that simply took in input and followed instructions to give an output which, when the instructions are fittingly complex, can give responses exactly as a human would without understanding the input or the data that is being output (Searle, 184-186)2. This example, which Searle could obviously conceive of, seems entirely possible, given a machine that is complex enough.

Searle also offers a damning critique of the notion that consciousness can be the result of computer programming alone, which seems to problematize this view. However, Searle also says that if a machine were designed to physically match the human mind, it would have consciousness (195)3. Therefore, Searle’s theory is incomplete somehow; there must be some way, perhaps not involving “programming” as Searle defines it, that could allow a machine to have consciousness. Since we do not know what this other element might be (a God-given soul? A dualistic mind, perhaps?), let alone how to generate it, we ought to be careful and give concern to machines, which just might have this unknown element in them.

Since it is possible but not certain that a machine could have consciousness, then, we should consider some particularly-intricate machines to be moral agents. It is difficult to say at what point of intelligence and programming complexity we should make such an ascription, but since the possibility that even highly-developed animals such as chimpanzees are still considered by some to lack consciousness (Colin), a conservative location to do so would be when a machine gains intelligence levels comparable the least-intelligent humans still known to have consciousness, which means, at least,  young children4.

This might appear to set the bar for consciousness and care too low, but every being with consciousness ought to be protected from suffering and servitude, regardless of intelligence level. As moral agents, humans have a duty to avoid harming other conscious entities that supersedes any gain that might be made on the backs of robots. If you make use of complex, intelligent robots for your own ends or destroy them without mercy, and it turns out that the robots were conscious, then you have committed a moral wrong that could not be met by any of the gains made by the (mis)use of said robots.

Footnotes

1 – I am equating ‘think’ with ‘have consciousness’ here, but I do not see this as a problem. One can only attribute thought to a conscious being, since the number calculations of, say, a modern computer do not count as thought, but programming. In other words, thought is an element of consciousness.

2 –Searle was not specifically arguing for the possibility or conceivability of zombies in his paper, but the argument he makes also seems to support such a claim.

3 – Searle specifically claims that the robot would have intentionality, but as with footnote 1, I consider intentionality to be a part of consciousness. Intentionality could not be ascribed to a being if it did not have consciousness, since one cannot have intent without awareness of one’s actions

4 – People with mental handicaps may, in some cases, have less intelligence than children while still retaining consciousness. As well, it may turn out that certain apes do, in fact, have conscious experience. In any case, the bar for determining when to protect machines for the possibility of having rights should be set at the lowest intelligence level at which we can guarantee that at least one being possesses conscious mental experience, since it is possible that other beings with equal intelligence may have consciousness then, as well.

 

Works Cited

“Solipsism.” Encyclopædia Britannica. Encyclopædia Britannica Online. Encyclopædia Britannica Inc., 2012. Web. 29 Jan. 2012. <http://www.britannica.com/EBchecked/topic/553426/solipsism>.

Allen, Colin, “Animal Consciousness”, The Stanford Encyclopedia of Philosophy (Winter 2011 Edition), Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/win2011/entries/consciousness-animal/>.

Kirk, Robert, “Zombies”, The Stanford Encyclopedia of Philosophy (Spring 2012 Edition), Edward N. Zalta (ed.), forthcoming URL = <http://plato.stanford.edu/archives/spr2012/entries/zombies/>.

Turing, A.M. Computing Machinery and Intelligence. Mind 59, 1950. 433-460 <http://loebner.net/Prizef/TuringArticle.html>

Vaughn, Lewis, and Chris MacDonald. The Power of Critical Thinking. 2nd ed. Don Mills, Ont.: Oxford UP, 2010. Print.

The Wisdom of Repugnance and the Uncanny Valley

Posted on by
Reply

I.  The “Wisdom of Repugnance”

Definition: The Wisdom of Repugnance (hereafter WoR) is defined as a kind of moral sentiment that can be felt resisting particular technological advancements for reasons that are not easily elucidated in rational terms.

In recent years, the so-called “Yuck Factor” or “Wisdom of Repugnance” has had its proponents. Bioethicist and Chair of George Bush’s President’s Council on Bioethics, Leon Kass invokes WoR in his argument against human cloning, appealing to the purportedly common sentiment that it is “’Offensive.’ ‘Grotesque.’ ‘Revolting.’ ‘Repugnant.’ ‘Repulsive.’”, etc. (Kass).  The sense of revulsion is, Kass argues, not to be ignored in philosophical debate as it stems from a deep wisdom that is simply beyond one’s ability to effectively communicate.  In the same way that it would be difficult to communicate the moral wrongness of something like incest or rape, Kass argues, it is difficult to fully capture the moral wrongness of something like human cloning in rational terms (Kass).  Despite the fact that the WoR isn’t an actual argument (as Kass rightly admits), it can serve as an indicator of a moral wrongdoing and a further rational elucidation of this revulsion can be useful in the moral analysis of a human activity.  Whether WoR is an appropriate tool for use in philosophical debate and its implications for the field of roboethics will be the focus of this post.

Bioethicist Leon Kass

Bioethicist Leon Kass

II. WoR and the Response from Relativism

Human sentiment is a dangerously relativistic device, deeply connected to our individual experiences and habituations.  A person who is raised in a Southern Baptist home in rural Alabama will have a different set of sentiments and visceral responses regarding morally questionable activities than a person raised in, say, a more liberal setting.  The allowance of such accidental features as influencing factors in a philosophical debate is, I take it, impermissible and runs the risk of falling into relativism.  Taking WoR as a starting point for a moral investigation, rational justifications of a sentimental position are reduced to mere reinforcements of the accidental position toward which the individual is predisposed due to their past experiences.

To be fair, Kass believes that he appeals to something so fundamental in the human experience that morally significant variations of the sentiment are unlikely, but this does not appear to be the case.  His own article examines the objections from pro-cloning scientists who, presumably, do not share in this supposedly fundamental repugnance.  Even if Kass attempts to claim that their inability to perceive the inherent offensiveness of their activities is due to a lack of the wisdom that he is privy to, this leads to the worry that he is simply mistaking those accidental features (such as the dispositions habituated in him by his past experiences) for an enlightened sense made accessible to him through his theological or personal commitments.  At this point, he can maintain that those commitments, despite being habituated by upbringing, are morally salient (and run the risk of falling into the relativism expressed in the above objection) or find a rational grounding for his belief (thus making the WoR dispensable in favour of more rationally constructed arguments).

III. The “Wisdom of Repugnance” and Roboethics

The threat WoR poses to a rational understanding of the potential of robotics is significant.  The familiar concept of the uncanny valley – a point at which the resemblance a robot bears to humans is close enough to be easily recognizable as an attempt at mimicking human traits while far enough away that it causes a sense of revulsion or fear – ought not to be given moral salience with respect to the place of robots in the world.

Everyone’s Favourite Uncanny Valley Photo

As robots fall into the uncanny valley, it might be tempting for some to appeal to some form of WoR to justify an argument that goes against their interests. This, I claim, should be an impermissible approach to the debate.

There are at least two reasons to deny WoR in roboethics debates:

(1)    Relativism objection

(2)    Aesthetics and Moral Salience

(1)    As expressed in section II of this post, basing a moral argument on a personal sentiment is poor form and potentially dangerous.  The personal feelings of individuals should not have an influence on the moral status of an activity and are certainly not a starting point for moral inquiry.  Personal feelings, it would seem, find their place only after the analysis of the activity is complete and the individual has rational grounding for their sentiment.  Approaching the topic from the other direction, from sentiment to rational elucidation, only serves to entrench interlocutors in the positions that they took as a result of their upbringing and promotes the building of an ad hoc rational framework to reinforce that bias.

(2)    The repulsiveness of certain robots, even if admitted, carries no weight regarding the arguments about their role in society and the rights they are granted as rational beings.  Whether they are, indeed, rational beings, whether more (or less) than reason is required to be a rights-bearer, whether exploitation of a machine built for a particular task is even possible are all legitimate moral questions we should be asking about robotics.  The feelings caused by the aesthetics of a robot is not a legitimate starting point for such a discussion, as an application of WoR would attest.  In the same way that the colour of a person’s skin or the features of their face are not morally salient details, the uncanny valley into which so many of today’s robots fall is not a morally salient fact that can be used in the discussion, despite whatever revulsion it causes.

I take these two reasons to be sufficient to deny the use of the WoR in philosophical debate regarding the status and role of robots in society.  WoR finds its grounding in subjective sentiments that are either dangerously relativistic or simply ineffective, and so WoR, at best, adds unnecessary complexity for comparatively little benefit.

Works Cited

Kass, L. R. (1997). The Wisdom of Repugnance. New Republic , 216 (22).