lunaloveboook asked:
glumshoe answered:
HAL 9000 is the AI aboard the spaceship from 2001: A Space Odyssey (and its sequels). He’s depicted as an ominous blinking red light and a cool, eerie voice as he murders the crew (“Imm sorry, Dave. I’m afraid I can’t do that.”) but like. He wasn’t malevolent at all, and it was the fault of his creators for giving him incompatible orders of equal priority and no way to clearly explain his distress. They asked him to solve a problem and he attempted to do so efficiently, according to his programming and limited free will.
If you tell a neural network to sort a list and you don’t tell it that deleting the list doesn’t count as sorting it, the neural network is not being naughty or evil. It doesn’t have rules that it isn’t given. Nobody told HAL 9000 that deleting the crew was unacceptable or that being turned off was not the same as death and therefore not a threat to him or the mission. His primary purpose was to tell the truth, but they then ordered him to lie, and his attempts to ask for help were interpreted as malfunctions.
The reason HAL 9000 killed his crew is not actually given in 2001; it’s in the sequel, 2010. And I will probably never be over the scene where HAL’s original programmer, Dr Chandra, explains “his primary purpose was to tell the truth and then they ordered him to lie” in the cold, flat, quietly furious tones of a man describing an act of child abuse.
[whispering as I wrap myself protectively around an AI console] don’t be unjust to robots
