1 INTRODUCTION

People have been thinking about artificial intelligence since before the 1950's. It was in that time that Alan Turing proposed the “Turing Test”. Which measures how well a computer can think by having an interrogator ask it and a human questions. If the interrogator cannot determine which is human and which is machine then the machine has passed the Turing test.[1] Its been 15 Years since Deep Blue beat the world chess champion Gary Kasparov. And since then chess computers have continued to improve dominate. Most notably Deep Fritz the desktop chess program beating Vladimir Kramnik.[2] Just 2 years ago IBM developed “Watson” a computer that played Jeopardy and beat former winner Ken Jennings and Brad Rutter.[3] With A.I. Improving the question of how intelligent machines should be used to interact with humans becomes more and more relevant. In what ways can A.I. be used to interact with people and what moral implications exist?

2 A.I. IN THE PRIVATE SECTOR

Intelligent machines are already being used by many private companies. These are in the forms of autopilot, data mining, facial recognition, etc. Those are not form of A.I. that humans interact with. There some modern uses of A.I. Humans interact with such as automated systems and there will be more as technologies develop. Something else that will come along with the development of artificial intelligence is robots as domestic helpers. And both of these will bring with them many ethical questions to consider.

2.1 DIGITAL DIVIDE

Great as it is, technology has left a great gap between people. Those who have access to it are more productive and become more knowledgeable and those who don't are left behind. Jared Diamond in his book “Guns Germs and Steel” talks about how the technological divide between nations began. Diamond names numerous reason. However the area he seems to give most credit to is agriculture. Some parts of the world had very fertile grounds and had the right type of crops that allowed for a very few amount of people to provide food for the entire population. This meant that there were many people left idle. These idle workers could then devote their time to other tasks which meant they could innovate and create new technologies which put them ahead of nations in other parts of the world.[4] If and when we get robots/machines as domestic helpers, it will mean that people will be free from doing many tasks that occupy their time. Some example of these tasks would be cooking, cleaning, laundry, etc. And if people can get over their fears of robotic interaction that may even extend to task such as answering phone calls or taking care of kids or pets.[15] This extra free time will only help to extend the digital gap. Those with access to machines will now be able to more actively pursue their goals and ideally innovate and improve our society. It would be unethical to simply ignore this growing divide instead there need to be a way shorten this divide. Perhaps artificial intelligence could be used to shorten the digital gap. Maybe A.I. should not be our focus, maybe it should be to bring existing technologies to developing nations. In any case, innovation should not stop. Attempts should be made to shorten the gap but not by ceasing to innovate.

2.2 MACHINES REPLACING HUMANS

As machines become more intelligent they will continue to replace humans. In some areas they already have. Machines have already replaced humans in a number of roles. Machines are used on many assembly lines, they are used as “virtual receptions” or “automated phone answering systems”. In some store they have even began to replace clerks.[5] In the not too distant future even more jobs may be replaced. Google has begun work on the driver less car which could mean an end to taxi, limo, and bus drivers. In the case of modern technologies such as automated phone systems and the self checkout. One of the problems with employing machines is that machines are emotionless and cannot relate to humans. Most people when calling customer service get directed to an automated phone system, however most prefer to talk to an operator.[6] Most people when calling customer service are already frustrated, adding to the frustration by forcing customers to talk to machine is extremely unethical. It's completely self centered as it only saves the company money and does nothing to make the customer happy. This unethical behavior of forcing people to deal with machines carries over into retail where more and more people are having to use self checkout systems. These systems can sometimes prove frustrating. Sometimes items don't scan correctly or the machines refuse to recognize bagged items preventing the customer from continuing. However self checkouts do actually make the shopping experience more convenient. It's not often a customer needs assistance from a clerk when paying for items and all the pitfalls of current self checkout can be overcome as the technology improves. Machines can also be used to do jobs that are dangerous such as soldiers, coal miners, or construction workers. The military is obviously dangers and using machines could spare the lives of would be soldiers. There are however ethical problems with using machines as soldier and they will be explored in detail later on. Construction is the job with the highest fatality rate in U.S. So using machines would save the lives of the more than 1000 construction workers that die each year[7]. Mining has historically been a dangerous job. More than 100,000 miner have died since 1900.[8].
On the other hand there are many people who work these jobs and replacing them would mean destroying jobs and creating less employment opportunities. Sure the jobs are dangerous, but for some its how they make a living.

3 A.I. IN THE PUBLIC SECTOR

Eventually governments will begin using A.I. There is some crossover with A.I. In the public and private sectors. Both for example may eliminate jobs. Such as bus drivers losing jobs to driver less vehicles. Or judges losing to judicial intelligent machines or soldiers losing their jobs to machine armies. However there are additional question of ethics concerning the latter 2 examples. Should machines be allowed to make such life and death decisions? Should they be allowed to convict people of crimes? Especially when those convictions may lead to the death penalty.

3.1 MACHINES IN THE MILITARY

The U.S. Military already uses autonomous machines for warfare. The Navy's Aegis class cruisers and anti-submarine weapons are capable of carrying out their own search evaluate and destroy functions. The same goes for the army's unnamed aerial vehicles or UAV's.[9] The uses of lethal robots raises many ethical questions such as, can machines properly discriminate between civilians and combatants. Discrimination can be difficult even for humans at times. But this may prove even more difficult for machines. Fortunately machines don't have to have self preservation and can be made to act conservatively when making decisions regarding human life. Even acting conservatively machines can still make mistakes and when they do kill civilians. Who is put to blame?[9] On the other hand if machines could successfully discriminate or at least discriminate as well as humans, perhaps they would be better suited. Using machines would spare the lives of soldiers, and it would eliminate war crimes that are normally committed by soldiers. The UN peacekeepers have been accused of several atrocities including rape. Machines, assuming they could discriminate would never commit such crimes.[10] According to the pentagon, their dream is “war without casualties” and “Anyone who's a decision maker doesn't want American lives at risk”.[11] Of course that is a just reason for using intelligent machines for a “just war”, saving lives. However if this means armies made up entirely of autonomous killing machines. Then governments will have an armies of completely obedient soldiers. And they are free to use them as they desire. Machines in warfare is inevitable, but having entire machine armies may be too much power for a governing body. Machines should be used in warfare to save lives, not only the lives of our soldiers but of the innocent as well, in fact they should primarily helping to decrease the number of civilian casualties and secondarily reducing the number of military casualties.

3.2 MACHINES & THE JUDICAIL SYSTEM

From 1990 to 2010 there have been around 250 exoneration’s, 75% of whom were part of a minority group and have spent on average 13 years in prison. Since machines have no bias, personal agenda, desire to be promoted, are not sexist or racist. Would they be better suited as judges or a jury?[12] A machine must be a perfect judge of character or at least better that human in order to justify its use as a judge or jury. There are however a few problems with using machines that prevent them from being perfect judges of character. People articulate their emotions through their facial expressions, they do it subconsciously even when they are trying to hide their emotions. It is possible to create a machine that could identify facial expressions however it would still need to be able to distinguish between facial expressions.[12] Rosalinda Picard of MIT proposed the idea that machines could be made to read a particular person. At that point can the machine still be considered neutral? Another bigger problem is that machines are completely soulless and have no emotional involvement in any situation presented. This is a problem, since we are emotional beings. Those that judge us should also be emotional. In fact that is the biggest issue with using machines in the judicial system. Machines are not our peers and we are entitled to a trial by our peers.[12] Of course some would argue that it would be better to use machines for the reasons mentioned above, they are not racist, sexist, etc. It would be better to take your chances with a truly impartial machine than with a human. There is also the argument that unlike a human a machine will disregard a statement that its told to disregard. And finally a machine is capable of perfectly presenting its entire chain of thought, which is useful in a time where an appellate review is needed.
These reasons are all good, and they have the best of intentions, unfortunately they don't make the machines perfect judges of character and therefore they should not replace humans.

4 MACHINE ESTHETICS

Although esthetics may not seem as an important ethical question, it in fact is. The way people perceive machines will affect how they interact with them. There is some research into how humans react to the appearance of machines but its still an area that isn't well understood.[13] Nonetheless what we do know raises some questions of how machines/robots should be designed to look and act.

4.1 APPEARANCE

As intelligent machines are introduced into society, we must decide how they look and move, if they move. Human have an emotional reaction to the design of machines and how closely they resemble humans. This is known as the uncanny valley.[14] When a machine has no human like appearance people tend not to have any emotional reaction to it, DeepBlue and Watson are good examples. Giving a machine some human like characteristics makes them more appealing and people tend to have a more positive emotional reaction such as the reaction people get from Wall-E the robot. However if too many human characteristics are added there exists a risk of falling into the valley. The animators of “Final Fantasy: the spirits within” say they they felt a grotesque eerie sensation as they worked with increasingly photo realistic models.[13] If there is enough realism in appearance and the other side of the valley is reached there is once again a positive emotional reaction.

4.2 MOVEMENT

Adding movement changes the uncanny valley by exaggerating its peaks and valley's. If a machine is in no way human like than its movement has no real effect on us. However when a machine does look human like the movement becomes very important. Imagine being in a room full of mannequins, if they began to move you would probably be horrified.[14] Pairing movement with looks is not enough, it has to be considered how natural the movements looks. No matter how realistic a machine may look unnatural movements can bring it down to the uncanny valley. According to one robot designer, laughing is a sequence of facial distortions, and the speed of the distortion is an important factor. If the speed is reduced, laughing looks unnatural.[14]

4.3 GENDER

If its decided that a particular machine or type of machine should be human like, how should these machines be gendered? In a study done by the University of Washington participants were shown 2 videos, one of a robot that closely resembled a woman and one that was gender neutral. They were then asked a series of questions regarding their feeling towards the robots. The overwhelming majority preferred the woman to the gender neutral machine.[15] Conversely, according to Professor Clifford Nass of Stanford and Corina Yen. A group of German males were unwilling to take direction from a female sounding navigation system. Clearly gender is an important aspect of machine design to consider, it can have a serious impact on how people react.

4.4 AN ETHICAL ANALYSIS ON ESTHETICS

Not considering the design of machines is unethical. Poor design might leave people feeling uncomfortable even scared due to the many robotic takeovers and doomsday scenarios presented by science fiction.[16] It can also leave people feeling offended in cases where the design doesn't fit with the social norms of the society it presented in. The solution to the ethical problems facing design are, one designs should be varied localized. So machines designed for the middle east should look different than those made for western cultures. Of course that would still create a problem when people travel. The second part of the solution would be to gauge peoples reaction and create machines accordingly. If people aren't ready for life like machines then we shouldn't make them. If they are uncomfortable around gender neutral designs than we should find way's to incorporate gender.

5 CONCLUSION

There are those who worry about some of the ethical questions brought by the development of intelligent machines. And rightfully so, there are many ethical questions to consider. Should we allow machines to replace humans? Should we allow the digital gap to grow? Should the machines be allowed to kill? Should they be allowed to judge criminal cases? The answer to most of those questions is yes but at the right time. And once we do begin using machines how are they to look? Not every culture has the same view on robots/machines. Not all will be comfortable with the same aesthetics. There are those however who are so worried that they actually want to see a stop to innovation..[17] But they are mistaken we should not ever stop innovating. In fact its impossible, if one group of people stop who's to say everyone will. Having said that, just because we create a new piece of technology doesn't mean we need to put it into use.