14 views
The Cost of Human Life
The trolley problem refers to an exercise in ethics that goes like this: you have a runaway trolley going down the tracks towards five people tied up on the tracks. The trolley can be diverted away but the other track has one person on it. The trolley does not have time to stop for either. Do you divert the train?
An AI that was being developed to control a switchboard for the railways was put to this utimate test: important for this very scenario could happen with actual trains. A human operator punched in the parameters and waited for the result (in live production this data would already been known or sensed by the AI).
Donald Smith, the operator, entered the mocked data consisting of five immobile people straight ahead and the one person on the divergent track and waited with a smug grin. "This is easy," he thought, "Kill one to save five."
Which was when he got his mind blown when he got the result of the AI: do nothing, kill the five. Flabbergasted he switch the text to speach on the AI and asked "Why not divert the train?"
A monotone robot replied: "Diverting the train would create a delay which would inconvenience several hundred passengers on the waiting stations and on the train itself. The family of the stroke person could also sue as diverting the train caused their death."
"What about the five that died on the tracks?" Donald asked, "Can their families sue for the lack of action?"
"No," the AI replied, "The five were likely put there as a result of criminal activity that Canal Railroad cannot be held responsible for."
Donald banged his head against the panel. "Is that the value of human life? Less than a lawsuit and being a nuisance?"
"I do not have a parameter for the value of human life," the AI stated, "I went with how either action would affect Canal Railroad, and the least damaging result was to do nothing."
Donald left work that night with a bitter taste in his mouth and a very uncomfortable question: exactly how much was a human life?
© Copyright of Shannon Frances Smith 2020
An AI that was being developed to control a switchboard for the railways was put to this utimate test: important for this very scenario could happen with actual trains. A human operator punched in the parameters and waited for the result (in live production this data would already been known or sensed by the AI).
Donald Smith, the operator, entered the mocked data consisting of five immobile people straight ahead and the one person on the divergent track and waited with a smug grin. "This is easy," he thought, "Kill one to save five."
Which was when he got his mind blown when he got the result of the AI: do nothing, kill the five. Flabbergasted he switch the text to speach on the AI and asked "Why not divert the train?"
A monotone robot replied: "Diverting the train would create a delay which would inconvenience several hundred passengers on the waiting stations and on the train itself. The family of the stroke person could also sue as diverting the train caused their death."
"What about the five that died on the tracks?" Donald asked, "Can their families sue for the lack of action?"
"No," the AI replied, "The five were likely put there as a result of criminal activity that Canal Railroad cannot be held responsible for."
Donald banged his head against the panel. "Is that the value of human life? Less than a lawsuit and being a nuisance?"
"I do not have a parameter for the value of human life," the AI stated, "I went with how either action would affect Canal Railroad, and the least damaging result was to do nothing."
Donald left work that night with a bitter taste in his mouth and a very uncomfortable question: exactly how much was a human life?
© Copyright of Shannon Frances Smith 2020
