| Whenever a person wants to present themselves | | | | mimic the niceties of human interaction, without |
| as an industry expert, one credible approach is to | | | | requiring machines to actually understand the |
| paint a shining picture of future technology and | | | | nuances of what they are doing. Machines can |
| what people can expect from hopeful visions of | | | | repeat these actions because a person has |
| things to come. One potential that has long | | | | programmed them to execute the instructions |
| bothered me is the current general perception of | | | | that offer these results. If a person wants to |
| artificial intelligence technology. | | | | take the time to impress aspects of presenting |
| There are a few key concepts that are not often | | | | their own personality into a sequence of |
| included in the general discussion of creating | | | | mechanical instructions, computers can faithfully |
| machines that think and act like us. First, the | | | | repeat these processes when called upon to do |
| problem with artificial intelligence is that it is | | | | so. |
| artificial. Trying to create machines that work like | | | | In today's market place, most software |
| the human brain and its special creative properties | | | | developers do not add on the extra effort that is |
| has always seemed useless to me. We already | | | | required to make their applications seem more |
| have people to do all that. If we succeed in | | | | polite and conservatively friendly to the end users. |
| generating a system that is every bit as able as | | | | If the commercial appeal for doing this was more |
| the human brain to create and solve problems, | | | | apparent, more software vendors would race to |
| such an achievement will also result in the same | | | | jump onto this bandwagon. Since the consuming |
| limitations. | | | | public understands so little about how computers |
| There is no benefit in creating an artificial life form | | | | really work, many people seem to be nervous |
| that can surpass us to further degrade the value | | | | about machines that project a personality that is |
| of humanity. Creating machines to enhance and | | | | too human in the flavor of its interaction with |
| compliment the wonders of human thinking does | | | | people. A computer personality is only as good as |
| have many appealing benefits. One significant plus | | | | the creativity of its originator, which can be quite |
| to building artificially intelligent systems is the | | | | entertaining. For this reason, if computers with |
| benefit of the teaching process. Like people, | | | | personality are to gain ground in their appeal, |
| machines have to be taught what we want them | | | | friendlier system design should incorporate a |
| to learn, but unlike us, the methods used to | | | | partnering with end users themselves in building |
| imprint machine instructions can be accomplished | | | | and understanding how this artificial personality is |
| in a single pass. | | | | constructed. When a new direction is needed, a |
| Our brains allow us to selectively flush out | | | | person can incorporate that information into the |
| information we do not want to retain, and are | | | | process, and the machine learns this new aspect |
| geared for a learning process based on repetition | | | | as well. |
| to imprint a long term memory. Machines cannot | | | | People can teach a computer how to cover all |
| "forget" what they are taught unless they are | | | | contingencies that arise in accomplishing a given |
| damaged, reach their memory capacity, or they | | | | purpose for managing information. We do not |
| are specifically instructed to erase the information | | | | have to take ourselves out of the loop in training |
| they are tasked to retain. This makes machines | | | | computers how to work with people. The goal of |
| great candidates for performing all the tediously | | | | achieving the highest form of artificial intelligence, |
| repetitive tasks, and storing all the information we | | | | self-teaching computers, also reflects the highest |
| do not want to burden ourselves with absorbing. | | | | form of human laziness. My objective in design is |
| With a little creativity, computers can be adjusted | | | | to accomplish a system that will do the things I |
| to respond to people in ways that are more | | | | want it to do, without having to deal with |
| pleasing to the human experience, without the | | | | negotiating over what the system wants to do |
| need to actually replicate the processes that | | | | instead. This approach is already easier to achieve |
| comprise this experience. We can already teach | | | | than most people think, but requires consumer |
| machines to issue polite responses, offer helpful | | | | interest to become more prevalent. |
| hints, and walk us through learning processes that | | | | |