Elon Musk thinks that we are on the brink of a Robopocalypse. With machine learning making leaps and bounds, it is easy to understand his fear. We’ve had warnings enough since Asimov’s day that as artificial intelligence outpaces our own ability to learn, it will become increasingly hard for us to control it. In the same way that our motivations are foreign and incomprehensible to the dogs and cats that we keep as pets, an artificial intelligence would desire things that we cannot begin to understand.
When extrapolating the possible desires of a hyper-intelligent being, we always come up with the worst one possible: the idea that eradicating humanity is the only way to save the earth, or robots, or whatever great absolute the artificial intelligence has been taught to value. This is the fear that always lurks at the edge of technological advancement: how do we know our creation won’t turn on us like a super-powered rebellious teenager, annihilating human life in some tantrum? It is this same fear that lurks beneath the eternal proclamations of doom that each new social technology incurs: Is Social Media Ruining Our Youth? The idea that technology is stripping something vital away from our lives is constant.
Of course, this idea didn’t stem from nothing. There was a time in western history when science and technology was seen as an inherent part of human existence. But, in the 19th century there was a big movement to separate out science from the spiritual. It is in this movement that changed the emotional center of Frankenstein, to the colder scientific one of Asimov. It is when the idea of technology as cold and soul-less, driven by heartless rationality, began. That idea wasn’t accepted whole-heartedly. Whole disciplines argued for decades over the role of science in a caring human community. As the industrial age plowed forward however, and the image of technology as bright, hard chrome took over, that idea spread to encompass everything about how we think of technology.
The problem is that technology is dangerous - but for all the reasons that Mr. Musk seems to not see. We think of technology as functioning under pure logic, existing with a sharp perfection that is unforgiving of human existence. But, technology is our creation. It doesn’t form in a vacuum, but is shaped by the needs and ideas and biases of the humans who work on it. It is messy, full of the logical fallacies and shortcuts that we already use to shape our worlds.
In fact, this is already causing issues. The technology industry has a serious diversity problem - even large corporations like Google have difficulty hiring qualified women and people of color. Unfortunately this is often reflected in the products created: automatic soap dispensers that can’t sense the hands of black people, Google algorithms that leave CEO positions out of women's job searches, and face detection software that can’t recognize asian faces. All of these are symptoms of a larger problem: the biases and stereotypes of people are reflected and extrapolated in our technology.
That is the real danger of technology: the illusion of objectivity. In thinking that any artificial intelligence we create will be greater than ourselves, we ignore the ways in which we are shaping that intelligence. By thinking of algorithms as somehow “above discrimination” we risk creating an illusion that the biases they present are based in fact.
The fact of the matter is that artificial intelligences are, at the core, artificial. They are created - shaped into something from nothing. They are part of humanity's legacy. Rather than fearing that they will choose to replace us, we ought to be focusing on how to train them to surpass our human failings. After all, it is with their assistance that we will reach the future of peace and prosperity that we all dream of achieving.