In AI Landscape

I sometimes do ask myself “What would John Cage do?” — and it’s been happening with regard to ChatGPT, etc., lately as the discourse around AI grows and grows. Writing about AI has become a kind of OpEd version of the “grey goo” feared during the initial development of nanotechnology. Forgive me for adding a few hundred GMA-free* words to it.

I imagine Cage saying various things about AI circa mid-2023, and some of those things are in opposition to each other. The opposition is not representative of him being mercurial, but of me being unsure about where his opinion would potentially land. In this context, I reread some of his conversations with Morton Feldman, collected in the great book Radio Happenings, and also some of his lectures, in which he talked about computers, radio, tape splicing, and other equipment he employed.

The following two contrasting statements seem to me what might, perhaps, be the outcome of his consideration of the topic — though I say that with the emphasis on the word “perhaps.” These statements in quotes below are entirely inventions on my part. An English teacher of mine, back in 11th grade, introduced the word “perhaps” to us as a highly useful rhetorical device, and it has been a key tool for me ever since. How does one employ the word “perhaps” correctly? When uncertain, say “perhaps.” When not in doubt, definitely consider saying “perhaps.” 

These two diametrically opposed hypothetical responses are where I’ve currently found myself:

“ChatGPT is just a tool, a tool made by humans. I’ve worked with creative constraints for much of my life. I want to see what this tool can do. For what is an iteratively refined prompt other than a form of explorative constraint?”

Or perhaps:

“ChatGPT is a machine — not even a machine, but software running on a machine. Why should I care what a machine has to say if all it knows is what we’ve already told it?”

I’m not sure which direction Cage would have gone. It’s sort of like how on the one hand, Cage was exactly the wide-eyed thinker you’d want on a committee of notable humans if we were suddenly contacted by an alien world, while on the other, his response might likely be something along the lines of, “We can’t even speak with our cats yet, and we don’t know what’s at the bottom of the ocean, so why do we want to leave the planet so soon?” Of course, before deciding one way or another, he’d ask if this alien world is home to a previously undocumented species of mushrooms. 

Almost certainly, Cage would have come upon on a direction — an insight — about modern AI, circa 2023, that I can’t myself imagine. But with the limits of my imagination, that’s the forked path I currently find myself at the crossroads of. On the one hand, he was an inveterate tinkerer. On the other, he all but gave up on the concept of using music to “communicate,” which makes me wonder if yet another means of communication would be of interest in the first place. 

I’ll continue my dive back into Cage, and I’ll keep thinking about this.

*generatively mediated authorship