Skip to content

I think, therefore AI am?

April 10, 2023

A few more doodles about ChatGPT, the world’s favourite bot, or rather on the engine that sits behind it.

Its simulation of human intelligence is impressive. It can be useful as a granular Google or an interactive Wikipedia. It and its mates will therefore change the world by making the world’s digital knowledge easily available without our requiring the intermediation of clunky old search engines. In that respect it represents as profound a step forward as we made when we moved from the gnomic utterances of the mainframe I operated in the early 1970s to Windows and WYSIWIG, and then to the internet.

It can be fun, in a nerdy kind of way. One day, it might be able to invent a joke that will make more than a small group of nerds laugh.

In the wrong hands, liberated from its guiding parameters, it could tell you (meaning any old would-be North Korea or Bin Laden) how to build an H-Bomb or to develop some nasty virus that reduces the world’s population by half.

But for now, until it develops some form of unhuman sentience, also known as the Singularity, we should remember that it’s a reflection of humanity and interact with it accordingly.

If it’s capable of passing the New York Bar exam, then the Bar Association is asking the wrong questions. If it’s capable of writing an essay on Henry VIII that fools British examiners, then we need to create new measures by which to judge human attainment and potential.

If it’s able to sum up the experience of other humans, it still can’t create our own experience, even if its output can form part of the backdrop. While it may be able to help us make sense of the experience and knowledge at its disposal, we still have to rely on our own cognitive ability to make sense of what we see, hear and feel. Common sense is not necessarily our sense.

If it’s able to write a “new” Beethoven symphony or a Beatles song from the White Album era, then fine. We might well enjoy them. Whether it’s able to synthesise anything approaching the cascading shower of Amens from the final chorus of Handel’s Messiah, a musical miracle that relies so much on the temporary wiring in our brains induced by the preceding passages of the oratorio for its overwhelming effect, remains to be seen.

Without new creations by human beings to update its database, its simulations will eventually decline – to our jaded ears – into ever-decreasing circles of banality. Like muzak. Because music has one essential component that the Beatles and Beethoven could never anticipate or create: its meaning to us. We might fall in love to the accompaniment of a song made infinitely more piquant by our knowledge that its creator killed themselves or died of tuberculosis shortly after writing it. Will ChatGPT ever create “our song”? That depends on us, I guess. Us humans.

Not that any of this really matters if the human condition is changed in an instant. In a nuclear wasteland we might not care about or even remember the Beatles. We won’t need lawyers and we’ll educate our children to cope with the necessities of the times. We’ll still love and lose. Perhaps we’ll go back to Homeric epics, passed on by word of mouth, to satisfy our need to create. Or make music with sticks, stones and animal skins.

And perhaps, on the moon or in a remote corner of the earth that an apocalypse couldn’t touch, a descendant of ChatGPT will sit behind a blinking terminal, perpetually incurious as to why its human clients don’t come to it for answers any more.

But most importantly, not caring. Because nobody has or ever will invent a computer that cares as humans do.

Then again, what do I know? Perhaps, with apologies to Paul Simon, “I am just a poor bot, though my story’s seldom told….”

Leave a Comment

Leave a Reply