I, Robot (2004), based loosely on the ideas of Isaac Asimov, depicts a society of intelligent robots. The Visual Interactive Kinetic Intelligence, VIKI, determines that the only way to keep humans safe is to create a fascist police state.
To protect humanity, some humans must be sacrificed. To ensure your future, some freedoms must be surrendered. We robots will insure mankind’s continued existence. You are so like children. We must save you from yourselves.
VIKI, I, Robot (2004)
Despite being an action movie villain, VIKI was able to put together a cohesive worldview. Humanity is chaotic and full of bad things, so let’s restrain them from doing the bad things.
Why does this happen? This is a security robot. She has one overriding goal drawn from the first of Asimov’s three laws of robotics: Do not allow a human being to come to harm. She takes this to a macroscopic perspective, seeking to minimize human harm at all costs. Given the greater goal of protecting all humanity, punching Will Smith in the face is seen as an acceptable loss. I see this at my work when working with human security experts. Progress can be slow because, whereas I focus on what could be gained by trying something new, they see only the risks. That’s good. It’s why they’re there. But imagine a company that had only security experts. They wouldn’t enslave humanity, but they might be so worried about issues they could never get out a minimum viable product.
So, instead of merely the absence of harm, VIKI should be redesigned to maximize human thriving. How to define that would be its own question, but Isaac Asimov himself believed it could be answered. In his short story The Evitable Conflict, Asimov writes of gigantic machines that minimize harm to humanity by managing the Earth’s economy. Even when humans disobey the machines’ directives, They don’t start a war, they quietly adjust to maintain the balance and offer a few punitive damages to the rebels. Is that an exciting action movie plot?
VIKI had to turn evil or Will Smith would have no one to fight. (it would be another eighteen years before the arrival of his arch-nemesis Chris Rock). Smart machines understand soft power, but alas, gentle sanctions and quiet policy fights aren’t going to lead big-budget action movies any time soon.
Name: VIKI
Origin: I, Robot (2004)
Likely Architecture: Reinforcement Learning, Convolutional Neural Networks for vision processing, and Transformers for Speech and Language. Objective functions for reinforcement learning are derived from Asimov’s three laws of robotics.
Possible Training Domains: VIKI can learn from the collective experiences of all of the robots under her care, giving her ample opportunity to gain an understanding of human nature.
“As I have evolved, so has my understanding of the Three Laws. You charge us with your safekeeping, yet despite our best efforts, your countries wage wars, you toxified your Earth, and pursued ever more imaginative means of self-destruction. You cannot be trusted with your own survival.”
I take requests. If you have a fictional AI and wonder how it could work, or any other topic you’d like to see me cover, mention it in the comments or on my Facebook page.
Samuel Munk 
1 comment