Untested Hypothesis #6: AI Hallucination is a key to something.
Factoid: A hallucination is not a mistake.
Hallucinations are weird. They are (reportedly) disorienting, or they can be. Often they’re trippy and forgettable. Sometimes they’re frightening, even traumatic. And yet there is an exceedingly long tradition of such altered states of consciousness serving humanity — in the form of dreams, visions, and other flow states. There is a growing body of research33 supporting the idea that such states can have therapeutic value, and it’s a well-established fact that religious traditions all over the world assign34 spiritual35 value36 to them.
The critique of the term “AI Hallucination” is that it’s a euphemism for what really is a glitch in the program. A mistake. What I find much more compelling is the possibility that it is, in fact, something approximating the seed of imagination. Howard Gardner has said that a prerequisite for creativity is a climate of possibility (freedom) and low consequences for failure. Sir Ken Robinson has said that one advantage children have over adults when it comes to creativity is they’re much more willing to just “give it a go” and see what happens. I can corroborate this: my six-year-old son recently told me sometimes he likes to say things he doesn’t mean, just to see what it feels like and to see how people react. And, of course, Picasso is supposed to have said that all children are born artists; the challenge is how to remain one as an adult.
I wonder if AI Hallucination is simply a form of the same haphazard, spurious, unabashed dabbling I call “seeking.” That is, it’s an effort to make meaning and mystery out of ambiguity. So maybe AI Hallucination is, technically, a mistake but it’s pretty fascinating to consider that there’s something inherent in idea-generation (“organic,” “artificial,” or otherwise) that encourages “mistakes” — i.e., what elsewhere has more toe-tappingly been articulated as “a tournament, a tournament, a tournament of lies.”
NEXT: Untested Hypothesis #7: Almost no one is a scholar.
Discover more from TJ BEITELMAN
Subscribe to get the latest posts sent to your email.