Comments
Sigurjon Helgi Kristjansson - 26 May 2018
A few decades ago, a Dr. asked his secretary to enter a few details onto a computer and assess if it was a computer or a real person she was dealing with. After only about 5 questions, she asked him if "they" could be alone. The questions were in such a friendly manner (unlike many Dr.s today, who lack bedside manner or go O.T.T.), that it made his secretary feel "comfortable" with the computer, and inputting details. A later improvement on this test, was a program that asked questions, and if data was lacking, would bring up comments and questions along the lines of: "Interesting, tell me more?" After only 5 questions, the computer gave output on another terminal, as to what the possibilities were, one of which the Dr.s thought was absurd. After 10 questions the computer had narrowed it down to about 5 possibilities, incl. the absurd one. After 20 questions, the computer made a definite diagnosis of the absurd answer. Because it was a rather rare occurrence, Dr.s dismissed it, but the computer didn't have emotions, and logic dictated that this was the most likely cause for the ailment. The computer was correct. If programmed with enough data, human error can be omitted, as can cases of misdiagnosis and unnecessary invasive examinations and prescription of expensive medication, which can cause a great deal of expenditure for the NHS as well as the patients, not to mention the emotional trauma and physical and emotional discomfort.