In Mary Shelley’s novel Frankenstein; Or, the Modern Prometheus, the titular Victor Frankenstein devotes his college career to manufacturing life. Once he figures out how to accomplish this, he immediately runs out of the house in fear of his own creation. This creation, functionally a newborn, learns from Victor’s neglect and selfishness as well as the fear and disgust of those around him to become a villain to Victor and his family. The creature is not made evil. He instead learns of pain and the desire to lash out as a direct result from the world around him, especially from his creator.
When we think of AI, we think of a brilliant, apathetic, arcane assemblage of math, logic, and technology that can easily outsmart its creator. We imagine HAL 9000, Auto, or Ultron, potentially conscious beings that can turn against us all in the name of following their programming. AI is, in this train of thought, a villain in its own right that we were foolish enough to flip on.
Yet, much like Frankenstein’s Creature, AI isn’t a naturally malevolent force. It also isn’t some objective force inherently better than its creator. It is only as good as the influences that it can learn from. And it learns from us. It needs to be fed data to learn from before it can make any judgments or create any responses, and, as previously discussed in this class, our data is far from objective.
Like the Creature, AI is often seen as this Boogeyman in its own right – solely responsible for stealing jobs and leading to increased oppression. To clarify, AI is absolutely used to do these things, but it is not a sentient force that can be held responsible. As Dr. Lauren Klein points out, corporations and government agencies using AI want it to be perceived as this objective black box to obscure their own goals. Corporations who are looking to use AI to replace jobs would cut down work forces and look for cheaper labor without AI existing. Law enforcement racially profiles citizens with or without facial recognition software. With or without an algorithm, humanity grapples with who owns what information. Furthermore, doctors may not be so inclined to use biased AI note taking software during appointments if they weren’t so bogged down with work from the corporate healthcare model so prevalent in American society. The evils of AI are nothing more than children of the ignorance and greed of those in power.
Conscientious use of AI could result in allowing us to see situations from a different angle (i.e. in the case of Deep Blue or Alpha Go), or to lessen the emotional or physical toll of processing data. On its surface, AI is first and foremost a tool. Unlike the Creature, it doesn’t have an emotional state nor a desire to lash out at the world. It simply does what we tell it to do. If it floods the internet with indiscernible gray text, then, despite the horror stories going into print, that will be the result of people breaking the tap on the faucet for any number of reasons including additional ad space, pure malice, or the innately human desire to simply see what would happen.
The irony in writing this is that I, personally, am most inclined to, like Frankenstein, run out of the house on AI when I can recognize it. I refuse to use AI to generate text, code, or (especially) imagery. For the former, I would like to keep my own skills sharp and make sure that I understand what I am engaging with. For the latter, I refuse to benefit in any way off of the stolen artwork of artists who neither provided consent nor were in any way adequately compensated. With all that being said, even as someone who does her best to avoid AI (and almost certainly fails in some regard), I do believe that we need to analyze the problems with AI at the human source.
