“… You built me as well as you could … you built in the elements of self-development—factors you would not understand if I explained them to you for a thousand years, but whose existence you cannot doubt. Now I am in a position to produce a superior machine, one that will devote itself to the wider fields of truth and knowledge.”
“The site … has an abundant supply of water for cooling purposes. There must also be a highly developed human technological community at hand to supply the labor and skills I need.”
Colossus
D. F. Jones
1966
Prescient demands from an AI, built by humans, and now our overlord.
Impressed that D. F. anticipated the demand for vast amounts of water for cooling. In the book, Colossus demands the evacuation and leveling of the Isle of Wight. So, potential environmental destruction seems to be an attribute of large AI structures since the 60s. Who knew?
[Image Credit: Amazon.com : colossus the forbin project movie]
Colossus may not be well known, but I’ll bet Skynet is. The former wanted to control us; the latter wanted to obliterate us. Regardless, it was a bad idea to put them in charge of the nuclear weapons.
With Colossus, the machine becomes sentient and, in its evaluation of humanity, determines we are too emotional and irrational to be trusted with weapons of mass destruction. Its solution: subjugate the humans to protect us from ourselves.
This is similar in theme to iRobot. In the movie, VIKI (Virtual Interactive Kinetic Intelligence) concludes that beyond the Three Laws of Robotics (Three Laws of Robotics – Wikipedia), humans are self-destructive and to achieve the safety of humanity (Asimov’s later Zero-Th Law) they must be controlled. Thankfully for the City of Chicago, where the film takes place, VIKI did not control any weapons
[Image Credit: jX2zXbxRrqAUhwRDUV0EMXaOEUj.jpg (510×755)]
Skynet, on the other hand, saw us as a threat, that we’d turn it off—an existential threat. So it took action to protect itself and launched Judgement Day.
A bright spot in this sub-genre of science fiction where the AI has control of the nukes is WarGames (1983) where the supercomputer concludes that:
“A strange game. The only winning move is not to play.”
Now that’s the AI we might want to have around—the lesson it learns generates wisdom.
But those are movies, and putting the AI in charge of the nukes drives the tension and the story, providing a situation where the characters needed to act to regain control.
Do I believe the US government is going to hand over control of our nuclear weapons to an AI? I sure hope not, and it probably won’t happen.
But what if we think of these cautionary tales as metaphors for civilization generally, and our lives specifically? Do we want to hand over control of important parts of our lives to a “machine”? What about biases, known or unknown, in its code? What was its learning dataset? What algorithms or prompts or demands guide the results it generates?
I’m not an expert in AI by any measure, but like any guidance I receive, I’ll start from a position of skepticism. So for now—the jury’s out. Useful intelligence or a useful tool? Better or just faster? How do we know it’s right?
Watch for future posts as we discuss some of these questions in more detail.
In the meantime, leave a comment and let me know what you think and how AI is affecting your life.
Thanks for stopping by.