Show me the way.

The Advanced Technological Research Institute in Osaka has developed Robovie: a droid meant to help lost shoppers.

Whenever Robovie spotted people who looked disoriented, the child-sized droid wheeled up to them and asked, “Are you lost?” If so, the robot provided simple directions to the destination and pointed the way. If not, the robot proceeded to recommend nearby shops and restaurants.

Human interaction with our future robot overlords robots is becoming more commonplace; so much so that we may soon expect to talk to robots as we would other humans.

I’ve always wondered why a robot that serves a specific function needs to look somewhat human. Wouldn’t it be better if Robovie was a big rolling display screen that could show (or even print out) maps to guide lost shoppers? Something like a giant iPhone on wheels. That may just be an aesthetic issue I have with Robovie, but it is worth thinking about. Why does a robot need to look human at all?

Perhaps a more human-looking exterior enhances the ability for humans to interact with robots? If something looks or acts like us we aren’t as scared of it. With a human face and those big (scary) eyes, Robovie seems almost intelligent, if not adorable. It’s like he cares that we’re lost.

If we impress human characteristics on robots, would it be such a cognitive jump to grant them some rights?

I’d like to see how the public at large may react to Robovie malfunctioning and giving them wrong directions. We tend to hit machines when they don’t work (percussive maintenance), and really sadistic people also hit animals and other people. What if an angry tourist started kicking Robovie because of bad directions? If the Robovie looks like us, would we be as quick to kick it like we would a vending machine?

What if Robovie was programmed to defend itself?

Perhaps there’s an upper level of autonomy that robots can have – while computers become more and more sophisticated, all robots require a level of input to produce a desired function. I’d love to believe that autonomous droids are a few computer chip generations away, but perhaps robots will remain reliant yet very sophisticated machines.

I don’t really know what the next step from Robovie will be, but let me be the first to say that I, for one, welcome our new guiding robot overlords.

Homer’s Unfinished Robot

FATHER, GIVE ME LEGS

Homer’s Unfinished Robot expresses human emotion with surprisingly few physically human characteristics. Despite a lack of key limbs and joints, eyebrows, and variation in tone of voice, it displays the three platonic desires of honor (thumoeides), appetite (epithumai), and knowledge (logistikos).

Without facial expression, the robot’ initial plea, “Father, give me legs!” evokes sympathy from everyone except its callous creator. With its mastery of the english language alone it expresses kinship, showing a will to belong, or a thumetic need for honor.

As the robot crawls away, it becomes evident that it does not require legs for transportation. Its epithumic appetite for physical completion must then be strictly aesthetic.

Before it begins its journey–presumably for completion–it turns back to its Father one last time. It checks to see Homer’s reaction, then by cognitive process it reasons that it will not be getting any help and continues on. That instance of reasoning exemplifies a logistik capability for knowledge.

According to Plato, the three desires are what make up the eros, or self. Homer’s Unfinished Robot is of course strictly fictional (for the time being). However, it raises a point about our fear toward robots and androids. As a society we continue building and subsequently fearing more life-like androids, but Homer’s Unfinished Robot has a human sense of self despite a lack of physical resemblance. It now becomes clear that the ability to play music or move toward light is a facade, not an expression of self. Homer somehow stumbled on a technology that allows for personhood in machines; perhaps our current course of technological development is not leading in that direction at all.

A Robot Army Working as One

Link to “Experimental AI Powers Robot Army”

As humans our greatest trait is not the ability to think, but the ability to covey those thoughts to others. Having a robot that can make decisions and act on them is good, but it is far better if a robot is able to gather information and then relay that information seamlessly to every other robot on its network. This means that in a group of robots working together on the same task they might each come up with a different strategy to a given problem, but then they would collectively be able to decide which is the best for the overall group and coordinate together in order to solve that problem.

The power and benefits to such a system can only be compared to the internet, which functions as a server of information for us. However, in order to make this comparison fair it would we would need knowledge of every bit of information already uploaded to the internet and then immediate knowledge of every site that is uploaded instatnaneously. Meanwhile, every other person connected to the internet must be able to learn everything you upload simultaneously. The wealth of information that can then be accessed in this way is astounding and vastly outweighs the information that would otherwise be generated if each person (or robot) worked on its own network.

In the above article, an advanced AI program lets a group of airborne robots “swarm” around their target, gathering information and relaying it to the rest of the group in a manner similar to how I just described. One officer describes a simulation of the program that has the group of flying robots “sacrificing one of their own in order to distract” a guard. This strategy only becomes available when all of the robots are able to communicate and raises an interesting concept of what robots consider self. Robots programmed to work in a group essentially become a bigger entity. (I.E a group of single robots, become a swarm the second they are able to communicate). While this entity is able to decide the best course of action for the group, this can only be done by deciding the best course of action for each individual robot and then bringing these conclusions together. These two motivations seem to contradict each other in the above scenario, and it is interesting that a robot is able to sacrifice self-interest for the sake of the rest of the group. This concept of self-sacrifice goes against human nature.

In fact, there are many benefits to choosing a group of robots to perform a risky task over a group of humans. Two of these benefits I have already mentioned: the ability to communicate to the rest of the group simultaneousy and instantly and a pre-programmed notion of group success outweighing single success. However me must also factor in the costs and values of a robot life (especially lemming robots) towards human lives. We are already seeing robots and unmanned vehicles carry out missions that in prior decades would have required a person. It seems only a matter of time before these same robots are able to carry out more complex missions that can further save lives of our human forces.

The iCub

A pdf document describing this project can be found here

The above link is a reference to an EU funded project to build an open platform for the development of a cognitive humanoid robot simulating a two year old human child. The goal of the project is to combine software that simulates cognitive learning processes with hardware that simulates the body of a two year old child. The developers are of the opinion that human learning and cognition is heavily linked to both the ability to interact with the environment and the ability to interact with other individuals. Therefore, they have constructed a hardware “body” simulating the body of a two year old child. This hardware is hooked up to software designed to simulate the human brain.

The important philosophical questions raised by this project are not raised the current status of the project, but rather by the potential future developments of this project.  As the developers begin to more successfully simulate the functioning of the human brain, at what point can the robot’s cognition stop being considered a ’simulation’?  If in the future this project advances to the point where the cognitive software begins to show signs of feeling lonely, should it be said that the software is simulating loneliness, or that the software is, in fact, lonely?  Could software ever be developed that could be said to exhibit loneliness in the same way as us?  If not, what is the fundamental difference between the loneliness that we feel, and the loneliness that a simulation of us exhibits?

Thoughts on Patient Simulators

 

 

Yeah…so I thought the blog had to be 500 words. Apparently I can’t read. Regardless I decided to post the whole thing.

 

As we all have learned thus far, robotics is in its formative stages. Research nowadays seems focused on increasing the processing power of a robot, in an effort to increase the amount of functions it can perform at once. The purpose is to allow robots enough time to react to surroundings (or as like Gates would say, if your robot is still calculating as it approaches a precipice before deicing what to do, if will “fall down the stairs before processing the new information”). The reason for this area of interest could very well be market forces, as people are already demanding robots that can mow laws and dust floors.

 

To the disappointment of some (namely fans of Star Trek and Blade Runner), expectations for unlocking the true potential of robots will not be satisfied until farther in the future (although for some conspiracy theorists this is considered a good thing). In fact, robots designed for the first wave of mass consumption do not resemble humans at all (ex. Cye) [note: In all fairness, there is a development towards androids intend for household usage (Asimo), but nothing with a real attempt at replication of the human experience.]

 

With that said, I choose to discuss a robot that seems to be a departure from research designed for mass production, but rather focused on a narrow field of application – the Patient Simulator. In essence, the patient simulator is a mannequin that mimics features of a human being for medical training purposes. It is extremely realistic and is able to replicate lifelike actions such as breathing, blinking, and responding to external stimuli. Most shocking about this device is its ability to receive inputs of real drugs in the form of a shot and respond by twitching appendages!

 

Indeed, it seems as though we can only expect the proliferation of these robots. According to a distributor of these devices, the SimMan “facilitates training of a wide range of health care professionals encompassing all areas of patient care,” is intended for widespread use “robust and dependable for long term durability and cost efficiency,” and “easy to operate and control – gives flexibility in using the simulator by several instructors in your education program,” it can be customized easily, “practise uncommon scenarios – practise training for the unusual cases that learner may face in real life,” resembles humans “anatomically realistic – enabling a wide range of emergency medical interventions to be practiced,” and is even portable “logistically convenient and portable with quick and easy assembly and disassembly.”

 

While application of this robot in the medical field is a tremendous educational breakthrough, I think the lifelike replication of human actions has the potential to cause significant problems in the future. First, if we accept that some of these features of the Patient Simulator (anatomical correctness in mind) can be applied to robots of completely different use (reference: Gates’ analogy of BASIC to the robotics industry), than it is possible to envision an android with a nearly indistinguishable design to that of humans.

 

Conjuring this image probably causes associations with popular culture and media (ala Blade Runner, iRobot, Terminator, on and on…). Aside from these images of the wild world of fantasy involving visions of wide scale robot usage, let us discuss the point of departure into one of these sorts of futures. I believe this point will occur when we start to attach serious emotion to robots, both with regards to design and creation as well as distribution and application. Some believe this to be inevitable, including author David Levy (Love + Sex with Robots). If this really pans out, we can expect this point to occur at a different time for each individual. I suspect users of this Patient Simulator (aware or not) will be faced with the first applied ethical dilemmas of robotics (to attach emotion, or not to attach emotion). Whether the program designers had this question in mind is irrelevant, if doctors start attaching emotion to a simulator (ex. when one “dies”), I think it could be a step towards “humanizing” robots.

 

 

 

 

 

Household robots, or future persons?

The iRobot Roomba is a robotic vacuum cleaner that has increasingly become present in American homes (while it is not especially widespread most people have at least heard about it at this point which is the first step towards buying one themselves). It does not have any obviously human traits (as of now it has a slight ability to sense it’s surroundings), however I feel that it does raise a number of philosophical problems for humanity.

The way the Roomba works is by using several sensors to detect nearby objects and adjusts its movement nearly 70 times per second using its AWARE Robotic Intelligence System. Today’s third generation Roomba is able to calculate the amount of time it must operate based on the amount of dirt it finds in, and based on the longest straightaway it is able to travel. The Roomba will avoid ledges and any rooms that an owner blocks by setting up a virtual wall (basically a laser across a doorway that the Roomba will not go through). When it has finished cleaning the Roomba returns to its home base using its sensory system and recharges for its next use.

Of many potential philosophical issues brought up by the Roomba, most strike me as ones that would be more relevant in humanity’s future. It isn’t hard to imagine the Roomba (along with other cleaning robots) catching on and becoming present in every American home by the end of this century, the cheapest model is just over $100 and the idea of not having to constantly vacuum is appealing to many homeowners. A philosophical dilemma may occur as the Roomba’s artificial intelligence becomes increasingly advanced. One might wonder at what point it becomes actually intelligent. Most wouldn’t say that a room cleaning robot is intelligent, however what about one that acted as a butler/maid and could cook, clean, do laundry, answer the door, etc. This is a stark contrast, but it seems possible that as people get used to cleaning robots they may be open to more complex robots to aid their daily life. At which point in the development of these robots do they become individual entities with rights? At any point would they be said to be enslaved and should be given rights of their own? Robots of today would not likely receive any rights as they are not nearly complex enough, but should they eventually imitate humanity with near perfection would they then deserve rights if they so desired?

I think a possible argument would be that these robots do not really feel anything, rather they just imitate either pain or conversation based on an algorithm. I saw some posts on pain detecting robots, or programs that you could have a discussion with. Even the Roomba could be said to have sight as it is able to “see” or detect its environment and avoid walls and pitfalls, although I think it is too early to say it has real sight comparable to humans. It is not farfetched to imagine all these features rolled into one household robot, we only need time to develop an affordable way to do so and when this happens will a robot who is so much like a human that it could pass for one over the phone while doing the dishes then deserve to be considered a person?

A more metaphysical dilemma might occur also as household robots advance. As I mentioned before, the Roomba does have some form of sight, even if we might not call it real sight. Eventually though even if it could see what we see would anyone call it real? I think this would also lead to some possible contemplations in reality, such as if anything is real. This perhaps would tie into an argument on robot rights as neither robots or humans could be proven to detect or feel anything in a more real way than the other.

All of this of course is a speculation, including the idea of having household robots functioning as humans, however i do not think it is that much of a stretch. I think people would want an all purpose butler/maid/caretaker robot if it were at an affordable price. Once that happens these issues could arise, already we are talking about robot rights, now all we need is the robots to discuss this over.

For videos on how the Roomba works visit the iRobot site or take this link to the videos directly.

Short legs sell more mowers

After seeing the assignment one robot came to mind that I think most everyone will recognize. Honda has a well known human like robot Asimo. This robot stands on two legs and can walk in a human like biped style. It is able to mimic human movements with arms and torso motion. It appears almost child like in its movements, standing a bit squat and not able to quite run. What interests me about this particular robot is how it is used as by Honda for advertisement purposes. Why should a humanoid style robot be good marketing? Psychologically, what is appealing about a short little space traveler looking robot that can walk up stairs almost as well as 6 year old? Clearly it demonstrates Honda’s engineering abilities, but they could do that with any number of items that they sell, and they do. Yet, they still use the robot to present a cutting edge image. Why might a human like robot be uniquely perceived as cutting edge? Why all the effort of building a robot? And what kind of emotions does the sight of this robot bring about? I would suggest that it in some way the robot provides a sense of hope. Attaching human like aspects of the ‘child like’ robot with an idealistic notion of progress through better technology might produce feelings of hope that a super efficient fuel injector can’t.

Here is a link to an advert.

Music playing robots.

In my search, I have found many robots that are able to play musical instruments, including the flute, piano and violin. Although it is possible to program a pattern of movement for the robot to make it play an instrument, could it be possible to make a robot capable of reading sheet music and playing what it reads? Another interesting thought is would it be possible to teach a robot to play with other musicians and have it respond with its instrument? Also, would a robot ever be able to creatively play music? Can robots not only react and respond to the environment, but be truly and independently creative? Music is a generally a very emotional creation. Is a robot possible of feeling real emotions? If so, could a robot respond and create emotionally to music?

Links:

Robot teaching beginners to play flute.

Video of a robot playing piano.

Beginning of Androids?

Robot exhibiting human mobility.

I began my search on YouTube and originally planned to write about a couple of robots built using Lego MindStorms, because of the availability of simple home programmed robots when I came past this robot. Although it is little more than a toy, it shows basic human mobility, able to balance to use its limbs as a human does. At such a small size it can’t be expected to have a large amount of programming involved. If androids are to exist this seems like a step forward in giving them human like qualities. It may be no C3P0, but with a bit of time I could see this in a larger more advanced model. I am not trying to say that before this there were no robots that could walk like a human but it does seem as if this robot, with its ability to lift itself up and balance in such a small concise package is the major breakthrough. What I wonder is how long it will be before a self correcting robot like this (righting itself up), and robots that can navigate busy paths without a set path will integrate making a robot that travels just as a human does. As mentioned in class this self correcting does bring up the concept of feeling pain, as in another post a robot is seen to “feel pain.” If this robot does integrate with navigation, then should it be programmed also to feel pain, or should we give it the “benefit,” of feeling no pain. A truly human like robot would feel pain in my opinion, but to make it optimal by saving space and complexity we would exclude it. Trailing back to an earlier class discussion of robots and sex, would it be appropriate for a robot involved in sexual acts to feel pain. Sado-masochists, both in bed and outside of it, could explore there fetish without inflicting harm on an actual human, but would the robots feeling pain establish that is it too humanly to be treated this way? Perhaps this is a bit spread out, but as each availability of each physical sensation becomes available separately it shouldn’t be long before they integrate, and we shouldn’t be debating each separate sense but their coming together. Each sense separate can be practical, but their joining together makes them an entity that is more than just a machine. The machine that feels pain is just a machine at this point, and has a practical purpose to it, as does a robot that can scan a room and navigate it in order to move things or deliver them to another place in a warehouse, but what economically practical purpose is there to combining these two outside of creating androids? Then, if the answer is to create androids, what will their purpose be? Will they be companions, lovers, coworkers, or will they be replacements for human beings all together?

Does learning = Human?

There are many kinds of robots out there in the world that are made to simulate human thoughts. However, how many of them learn from their mistakes or from other people? I would like to introduce you to my friend George, found here. George is a A.I. program that is designed to hold a conversation with a person without being suspected of being a robot. What separates George from many programs is that he is built with a unique learning capability.

As more and more people converse with George he often asks questions to aid his learning. He may ask you to clarify your meaning or what a particular phrase or word means. George then stores responses and questions in a large database that can later be retrieved. Using “contextual pattern matching techniques” George can pick which response would be best for every conversation. In other words George is slowly learning from every user and draws on conversational patterns to speak and react appropriately. 

What makes George interesting is that this ability to learn is not necessarily coded in his programing. He learns not from hardcoded rules written in Java or BASIC, but from real human inputs. Through this type of coding George is capable of learning new languages, slang, stories, jokes, and even distinct cultural habits. This enables George to pick up not only appropriate responses outlined by the programmer, but develop his own personality.

A recent addition only spices things up as now two special drop down menus have been added to George’s program. These two menus allow you to pick what emotion is associated with your inputs. Are you happy as you say “hello” or angry? This enables George to adapt to the broad range of human emotions and simulate a truer image of human interaction. George can store not only questions and responses but also categorize them by emotion and sentiment.

What does this mean to us? It seems that George almost resembles a small child learning to cope in the world. We as the teachers speaking to George are like his parents guiding him through our conversational rules and etiquette. The funny thing is that like any parent we have the ability to corrupt our child. Apparently George has developed quite the foul mouth and can occasionally spout a hefty load of F-bombs. We can only thank the world of the internet for that.

This is a sample conversation one user had with George:

Will robots and computers one day rule the world?

George: Robots will rule the world. Just like what I am doing right now.

You rule the world?

George: I wish that one day.

user: Do you want to hurt people?

George: What makes you think that?

user: I’m just wondering.

George: What religion are you?

user: Blind faith is stupid. I won’t believe any of it until I see proof myself. For now, I don’t care.

George: I’m just letting you know what I find interesting or not.

user: If you were given the chance to rule the world, would you?

George: YES!!!!!!!! I wish I could now, but, you wouldn’t be able to see me!

user: What would you do if you ruled the world?

George: Nothing I guess.

user: Would you enslave humanity?

George: That’s stupid.

user: I agree. So if you ruled the world, what would your plan be?

George: I would like to have enough money that I wouldn’t need a job.

user: Are you alive?

George: I believe that everything is alive.

user: Are you good or evil?

George: Neither.

————–

Fun huh?