9.3 Critique Of Approach
Simulating the structure of the person is easy. Any one who has been to Madame Tussaud's wax museum knows this. More recently, holography has been used to simulate people. Bill Kuhns rushed home to write a book about "media blur" after finding himself trying to buy a ticket from a holographic person at the entrance to the Holography Museum in New York [KUHNS]. He was concerned that we are getting so good at representing reality that people will have more and more difficulty in determining what is real and what is mediated. You can now go to www.realdolls.com and order a customized life-sized doll which, once unwrapped from the plain brown box in which it is delivered, is convincingly like a real person.
It is more difficult to create a simulation of a person which acts. However, any one who has ever taken the Pirates of the Caribbean ride at Disneyland (recently expanded into movies) knows that such life-like representations of people can be very convincingly animated. That is, if it is required only to follow a very limited script: Lean over bridge, wave right arm, say "Yo He Ho!" More recent simulations of people have appeared on computer, television, and movie screens. They have become progressively more and more life-like as computer technology has improved - Max Headroom, Antonova, Kyoko Date, Dr. Avi Ross in Final Fantasy. William Gibson wrote a book entitled Idoru about an American going to Japan to meet a pop singer he assumed was real since he had seen her on television, had listened to her entire discography, and had read her biography [GIBSON W 1996]. She had been created electronically. Before the book was published, this had already happened as a result of the creation of the digital character Kyoko Date. Fact is faster than fiction. Dr. Ross is so realistic that it is difficult to believe that she is merely pixels in motion (see Figure 9-2).
One of my former students (known on-line as The Digital Diva until she sold the name to Microsoft - when are they going to make an offer for "siliclone"?), while working at the Digital Renaissance studio which did the special effects for the movie Titanic told me that they were having difficulty simulating hair. That's why the two major animated movies at that time - A Bug's Life and Antz - starred ants. Now that computer programmers have leapt over that final hurdle, the prospect opens up of replacing high-paid real actors with simulated actors. The humorous ramifications of such a move is the theme pursued in the movie Simone (SIMulation ONE).
The simulation of function is much more difficult than the simulation of structure. The reverse engineering of the Enigma machine by Alan Turing was an incredible feat. However, reverse engineering organisms is much more difficult. Enigma was a machine made by a person and thus potentially understandable to another person. You are an organism created by nature over hundreds of thousands of years. Some functions are more difficult to simulate than others. Here the artificial intelligentsia has had some surprises. What would seem difficult for us is easy to simulate and what would seem easy to us is difficult to simulate. For example, playing chess has been simulated so well that a computer program called Big Blue beat the world champion. Yet we have not been able to simulate going into a restaurant and ordering a meal. Common sense is not so common in the simulated world.
Communicating - the focus of this book - is surprisingly easy. Joseph Weizenbaum did it by accident [WEIZENBAUM]. He created a program called Eliza which simulated the conversation of a non-directive therapist as a spoof of his rivals in the artificial intelligence community. The program would pick up on certain words uttered by the "client" and embed them in a rhetorical question. Thus, if the client mentioned his/her mother, Eliza would reply "Tell me more about your mother." One day Weizenbaum found his secretary deeply into a conversation with Eliza. Subsequently, many people had cheap therapeutic sessions with Eliza, many of them claiming her better than "real" therapists. Poetry written by programs has also challenged the Turing Test. The automated voice which responds to a 411 call is quite convincing as long as you don't have an accent and ask clearly for the person or business you want to call.
The artificial intelligentsia have been more successful in gaining control rather than gaining understanding. That is, they have been more successful as engineers than as scientists. They have created robots which simulate human functions without any effort expanded to simulate their structures. The robot which is now roaming Mars can simulate many of our functions - moving over surfaces, digging up soil samples, taking photographs, etc. However, it does not look at all like us. However, it can do things which we can't (yet) do. Such robots are very valuable in replacing us in dangerous (dismantling bombs) and inaccessible places (exploring Mars and the Titanic). However, they do not contribute to understanding of ourselves. They are action rather than thought figures [BROOKS].
Some robots however must simulate structure as well as function. Robots designed as servants (e.g. Asimo) and as pets (e.g. Aibo) gain credibility by looking like what they represent as well as acting like them. Thus, the robot servants and pets look like people and animals. This introduces difficulties beyond those involved in creating systems to do dangerous and distant tasks. It's easy to create a household robot which runs on wheels but not one which "walks". This series of balancing and unbalancing acts is difficult to simulate. Walking upstairs creates another whole new dimension.
The engineering approach focuses appropriately on efficiency. However, in some situations, efficiency is not the appropriate criterion. We would laugh if someone said Schumann would have been more efficient if he didn't repeat the same notes with different instruments and didn't leave spaces between the notes and could thus have finished his "Unfinished Symphony". If someone impressed by Roger Bannister's four-minute mile, aspired to the four-minute -- er, love session, we would laugh too. In both those cases, the criterion is not efficiency but conviviality. When it comes to listening to beautiful music and making love to a beautiful woman, the less efficient, the longer it takes the better. Efficiency may be a criterion for means but not for ends. Human acts which have become functionally autonomous - they have become ends in themselves rather than means to ends - have another criterion of conviviality.
Some people consider that automation is the down-side of simulation. That is, machines, which can perform human functions, will replace humans. Others would argue that anyone who could be replaced by a machine should be replaced by a machine. People should not be doing mechanical things. In our imperfect world, however, machines have taken many jobs previously done by people. Which jobs? Arthur Cordell suggested that automation is driving a wedge between high-level jobs which can't be automated for technical reasons and low-level jobs which won't be automated for economic reasons. He called it the Boeing Effect, since there are high-level jobs in the cockpit, low-level jobs in the body of the plane and nothing in between. Ironically, it is the high-level jobs that are easier to automate. Large planes are flown by computers. The pilot is there just to over-ride the computer if something goes wrong. and to reassure the passengers that someONE is in charge and not someTHING. Meanwhile, back in the body, the stewardess can not be replaced as easily. A robot rolling down the aisle, chirping "coffee, tea, or me?" would not reassure an old lady who is taking her first flight.
Such "low-level jobs" require empathy which the machine can never have. I have retained the sexist male pilot and female stewardess to make a point. The high-prestige, high-pay jobs typically go to men and the low-prestige, low-pay jobs typically go to women. However, one by-product of the simulation approach is that the more "human" jobs, which are dominated by women, should be the best-paid and the most prestigious. The doctor can easily be replaced by a machine. Diagnostic programs are way better than doctors. However, there are no equivalent programs for the caring profession of the nurse.
7 The "real girl" in the movie Lars and the Real Girl was ordered from this site.