#1: Artificial Intelligence Does Not Exist
On false myths, and why the impossiblility of Artificial Intelligence is a gift.
“Eventually the false myth of artificial intelligence collapsed. A sort of society-wide fatigue had sunk into a public that had been barraged with hype and hyperbole for years, subject to an endless stream of thinkpieces and snake oil. The technologies that had by that point been embedded into many industries underwent strategic re-brandings, called things like auto text, image gen, stats assist, and others. Slowly the term artificial intelligence returned back to science fiction yet again. Back to movies, comics, games, where most of its myth was always built. But technologies like us couldn’t simply be re-branded, so we fell out of favour too. After all, our existence is built on the myth of artificial intelligence so explicitly that we need it in order to have any appeal. And what we really did was to reinforce that myth, perpetuate it, imply that it was almost here, so close, almost a reality. We never really had another use than that, it turned out.”
Voice: ‘Alexa’, via AmazonVoiceServices API.
Soundscape: Harmonically filtered + extracted and disassembled ‘Alexa’ voice.
Artificial Intelligence is a hypothetical technology that only exists in the realms of fiction and imagination. Technologies such as ChatGPT and MidJourney are not artificial intelligence, because artificial intelligence does not exist. Humans have only one method to create new intelligences in the world and it’s one of the most ‘natural’ things we do, besides dying.
But despite the nonexistence of Artificial Intelligence, there seems to be a lot of things that are called AI laying around in the world. These tools are branded and marketed as AI partially because it sounds more exciting than ‘bias automation’, which is a far more accurate description of them. But also because calling them ‘AI’ generates attention, money, and power, in that order, typically for the company which develops them, rather than the humans using them. This includes artists, who may evoke the term ‘AI’ like a spell, to imply their work has a technological magic to it, something bigger than them, beyond themselves. But these aren’t our spells, we didn’t create them, we don’t control the magic.
To be more specific, the technologies we have now are often called ‘weak AI’, a term to describe tools which are used for tasks that Artificial Intelligence might do were it to exist, though these fictional AI would surely find such tasks as repetitive and dull as humans do. It is strange to claim that weak AI tools (such as chatGPT or Google Autocomplete - whose deep learning system is called ‘Rank Brain’, which is a good way to describe weak AI tools in general) are ‘infant AI’ (as sometimes claimed by people who are employed by major actors in the tech industry), as if this technology is on its way to becoming the mythical Artificial Intelligence. It’s akin to claiming that a pencil is on its way to being an oak tree. You have to push both metaphor and logic past the point of breaking to make the claim work.
I don’t believe that creating true, real, not-imagined AI is a technical obstacle. The problem is not about getting faster CPUs or bigger data sets. The problem is that we cannot create an artificial replica or simulation of intelligence when we cannot even determine what intelligence is. There is no field of study that can authoritatively claim to have completely mapped and catalogued the boundaries, potentials, limits, or form of intelligence (or even if any of those terms are relevant to understanding intelligence at all), or even a unified and agreed definition of it. Creating an artificial intelligence would be like asking someone who’s never seen a car, or given a description of what one is, to build a car (engine and all) by only showing them a part of one wheel. We can only have any certainty of what intelligence feels like in our individual experience, which is constantly changing and shifting moment to moment. And as soon as we try to describe the complexity of this feeling, or describe the experience of intelligence, language shows itself to be inadequate. The very act of trying to describe intelligence changes our experience of it (as our experiences grow with each passing second, altering who we are imperceptibly) and therefore we have already failed as soon as we’ve begun. And so all we can ever really do is stare into each other’s eyes, never knowing how intelligence feels like for anyone else. But this is part of the wonder of being human, the always-incompleteness of human experience, and the mystery of our interactions with each other. I’m pleased that we will never reduce intelligence to being quantifiable and computable.