The Sorcerer’s Apprentice, artificial intelligence, and things we fear.
TRANSCRIPT
About two weeks ago, I was reading stories to my son before bed; and since we happened to be staying the night at his grandparents’ house (my parents), I had scoured my childhood bookshelf to find a few good stories for the occasion. One of the selections was Disney’s version of The Sorcerer’s Apprentice, as in the book version of the animated short that was first featured in the 1940 film, Fantasia. In this version of the tale, Mickey Mouse plays the role of the pupil of a wise and learned magician, someone with the power to conjure images and enchant inanimate objects, as well as the power to make those conjurings disappear.
One day, the sorcerer needs to leave the workshop for business, and he sets Mickey the task of filling a large basin with water from the well outside. After a few trips with his two small pails, Mickey notices that the sorcerer’s book of magic has been left unlocked, and being already weary of the chore of lugging the buckets of water from the well, he wonders if he might use discover some supernatural method for finishing the task. Eyeing an ordinary broom in the corner, he recites an incantation to bring it to life; and as the broom sprouts arms and legs, he commands it to continue the work of hauling water to fill the basin.
Delighted with his accomplishment and tired from his work, Mickey falls asleep and dreams of being a masterful magician himself, standing atop a sea-side cliff and wielding command over the vast, immense waters. When he wakes, the workshop is flooded, as the bloom has no regard for the fact that the wash tub is already full and is now overflowing. Mickey tries to stop the broom - shouting new commands and utter other spells for the sorcerer’s book - but it is all in vain, and the broom continues pouring water into the ever-brimming basin.
Desperate for an end to the crisis and running outside with the broom to try and stop it from reentering the workshop, Mickey catches sight of an ax leaning against the wall near the door. With several blows with the ax, Mickey reduces the broom to mere splinters and breathes a sigh of relief, assuming that the ordeal has come to an end. However, the shards and splinters of the broom - still inhabited with the magic or the spirits that Mickey had summoned - begin to become animate, each and every one growing to a full-sized broom with two legs and two arms each carrying a water bucket. Soon the army of obedient brooms is pouring water into the workshop, and as the level of the flood works its way up the walls, Mickey begins to be sucked into a whirlpool.
It is, of course, at this final moment of crisis that the sorcerer returns from his business, looks upon the horrific scene and the mess that has been made of his workshop, and speaks the words that for years and years he has known by heart - dispelling every bit of the mayhem that Mickey had put into motion. And finally, as Mickey picks up the buckets to continue with his mundane task, the sorcerer gives him a gentle swat with the once again ordinary broom, hurrying him out the door.
When I finished reading, I was struck by the idea that the story made for a suitable allegory for the feverish rate at which Artificial Intelligence is developing. It’s certainly not a new idea. The out-of-control and potentially dangerous nature of the apprentice’s broom has been spun out in countless dystopian fictions centered on the fall of human society at the hands of a technological superintelligence. While films like The Terminator and The Matrix have captivated our imagination for decades, the theme of computer intelligence posing a threat to humanity has as of late made its way into our mainstream news headlines as well. In an open letter posted in March of this year, AI experts called for a pause in the training of powerful AI models so that the industry might have an interval of time to collectively consider how best to govern the continued expansion of this novel technology, ensuring that the potential for harm it presents is mitigated as best as possible.
The concern that AI could eventually outpace its creators’ ability to manage it also made its way into the congressional testimony given last week by Sam Altman – CEO of the company, OpenAI – as he called for close government regulation of the Artificial Intelligence industry. While initially avoiding details when expressing that, “If this technology goes wrong, it could go quite wrong…,” Altman eventually referred to the possibility of the technology “self-replicating” and “exfiltrating into the wild.” Thought experiments have described such occurrences - both in hypothetical instances in which an AI model was fulfilling its programmed task to the fullest, as well as pursuing novel ends of its own devising. Both for me evoke the image of the splinters of broom rising up into an army of hellbent aquarians.
Having never learned the origins of the story, The Sorcerer’s Apprentice, I was delighted when a quick online search told me that Disney’s animated short (as well as the orchestral music for which it was created) was inspired by a poem by Goethe, the same German writer who penned the text that inspired Beethoven’s ninth symphony. Perhaps Goethe was inspired by at least one of many old folktales with similar narratives; the most similar one I came across being the English story known as The Master and His Pupil, in which the pupil summons the demon, Beezelbub, who threatens to kill the young lad should he fail to set him task. After being ordered to water a flower within the master’s workshop, Beezelbub pours barrel after barrel into the room, flooding everything within.
In thinking on the rising water in these tales, I’m also reminded of the rising sea levels brought about by our changing climate, another out-of-hand situation caused by human conjurings. And it so happens that Goethe wrote his poem in 1797, just as the Industrial Revolution was about to spread from the United Kingdom into Germany and the surrounding countries - a singular cultural phantom splitting into many with each pouring buckets of carbon dioxide into the atmosphere, the unintended consequence of the unprecedented economic progress with which we had tasked it.
Who knows how old stories such as The Master and His Pupil are? And let’s not forget the many myths - told for millennia the world over - of the Great Flood that came as a response to the misguided, foolish, or ill-hearted actions and attitudes of humans. It seems that we have long been aware of our capabilities to summon forces beyond our control, as well as the hazards inherent in getting exactly what we ask for before thinking through the implications of the request. To the egoic mind that would love to think it has everything under control, this undomesticated, unconscious power is so numinous - so threatening and fearful - that we often deny it even exists, let alone that we have anything to do with it.
And so looking back to the issue of Artificial Intelligence, the question hangs in the air unanswered: Just how out of hand can things get, and will we even know when we’ve reached the point of no return? Remember that in our tale, there is a sorcerer who knows the words to put everything right again. In his testimony before congress, Mr. Altman made the suggestion that the government form a new regulatory agency to enforce strict safety protocols within the AI sector. In response, a congressman asked him if he would be qualified to head such an agency - an implied invitation that he politely declined. In this exchange, it seems that everyone is asking, “Who is the sorcerer here?”, and then looking around with expectant glances, as if the most authoritative party is about to step forward and tell everyone that it’s all going to be OK.
There’s a common theme running through all of this; it’s infused into the dystopian films; it pecks at us through the words of the open letter and the plethora of articles being written on the topic; and it drips from the walls of the congressional chamber where Altman delivered his exhortation to lawmakers. That theme is fear. We are all swimming in a flood “what if…?”, and the water just keeps getting deeper.
So what do we do about the fear?
The word fear comes from the Old English faere, meaning “peril, a sudden danger or attack.” Interestingly, the same Old English word also refers to a “path, journey, or expedition.” This is where we get the modern word fare, as in thoroughfare - a path to walk down. The Old English faere also comes from the Proto-Indo European word, per, which means “to try, to risk”, and also “to press forward” and “to move through.” So here we get the (perhaps) less than novel - but easy to ignore - notion that fear can (not always, but at times) indicate something we must move towards, move through, or walk along - even though it may involve some risk.
This last point is important to note. But what is even more interesting to me is that the Old English word, faere, indicates external circumstance that might be fear-inducing, whereas the general modern notion of fear is something that we experience internally. In other words, as time has progressed and language evolved, our understanding of fear has migrated from a place outside of us to somewhere inside of us.
This reminds me of the concept in depth psychology that describes how the unconscious parts of the psyche will project themselves - autonomously, without our knowing it - onto outer circumstances. Those who work in the field of depth psychology with an eye towards healing say that the unconscious parts of the mind do this so that they may come to be known by the conscious ego and eventually integrated into our personality. Like the word fear, here is another instance of things moving from our physical exterior into our psychic interior; however, they are things that originated within us in the first place.
In depth psychology, the unconscious mind is a force beyond our control, it acts autonomously, and it is powerful beyond our understanding. That sounds familiar. For those who study dreams and stories, the sea is often seen as a symbol of the unconscious mind. And as we talk now about deploying coastal defenses to protect our communities against rising sea levels; and as we hold congressional hearings to discuss how we might protect ourselves from computer intelligence run-amuck; I’m reminded of how we construct psychological defenses that partition our delicate egos from the flooding forcefulness, overwhelming vastness, and shear power of the unconscious realms of our very own psyche.
Perhaps you can see where I’m headed here.
Let’s conduct a different kind of thought experiment with AI. Here’s the question: What if our invention of computer intelligence is the manifestation of a projection of what psychologist Carl Jung called our Collective Unconscious? It would follow that this technology is just one of many ways that the depths of our own minds are trying to speak to us about the magnitude of power that we unknowingly wield. Not the power of the tools we wield, but power literally invested in our own capacity for consciousness. It’s fascinating to me that, at the same time that this AI boom is occuring, research in the fields of Consciousness and Neuroscience is also exploding. Mythic imagination might suggest that these studies are two sides of the same coin: one external, one internal; one reaching for new heights, and one exploring unknown depths.
In Goethe’s poem, when the sorcerer returns and dispels the spirit called into the broom by the apprentice, he says something to the effect of “it is only your old Master [the sorcerer] who shall call you forth again.” So, if, like Altman and congress, we are looking around the AI scene wondering who the sorcerer is going to be - the one who has the power to get things back in order after they’ve gone awry - perhaps we need to ask: Who is the old Master here? From where did this phantom originate?
And the answer, if you ask me, is all of us. Consciousness is a funny thing in that it does not observe the same personal boundaries that we like to pretend exist. Consciousness doesn’t stay locked up in the black box in which we’d like to believe that our AI models will obediently remain. Consciousness passes through walls, travels across oceans, and jumps into and out of the energetic melting pots that exist unseen people and among groups of people. You are the sorcerer. I am the sorcerer. Our children are the sorcerers. We are all a part of what is happening in this moment. And the old idea is that the more conscious and connected to the spirit of life that any one of us becomes, the more conscious and connected to the spirit of life we all become.
And so I’ll do something a little different today, and end with a question. For those of us wondering what to do in response to the possible AI revolution - those of us who don’t write computer code or legal code - certainly we can consider our use of the technology, and certainly we can write to our lawmakers and express our thoughts on the matter. But if, like me, you’re willing to play with the idea that the whole business of AI is perhaps a call to greater knowledge of the Self - the dark, powerful, unsettling parts of the Self - you might consider this question: When is it that you have demonstrated a power that you didn’t know you had? It could be some moment when you become uncharacteristically angry; or perhaps one of those delightfully uncanny coincidences when you thought about a friend you haven’t spoken with in years just before you receive a text message from them. When is it that you’ve had such an instance that was so surprising that it was even a little spooky or just plain frightening? That fear is an indicator. Some part of you wants to be known. Are you willing to take the risk to know it?
NOTES
The Sorcerer’s Apprentice by Goethe: https://en.wikipedia.org/wiki/The_Sorcerer%27s_Apprentice
Open letter from AI experts: https://futureoflife.org/open-letter/pause-giant-ai-experiments/
Sam Altman’s congressional testimony: https://www.youtube.com/live/TO0J2Yw7usM?feature=share
Photo by Maximalfocus on Unsplash