On: Who should own the knowledge that underpins AI technology?

Fifteenth of September, and the question before me concerns ownership of knowledge. The claim is that these new engines rest upon decades of work done by scholars who never sought a price for it.

Let me grant this provisionally. If a house stands upon a foundation laid by others, who owns the house? The mason, says the mason. The man who paid him, says the purchaser. But the mason’s apprentices learned their trade watching older hands, and those older hands learned from men now dust. Does the debt expire?

But what exactly does “own” mean here? When I know how to shape a sandal, the knowledge is in my body and no one can take it. Yet the shoemaker’s apprentice carries the same craft away and teaches his own son, and we do not call this theft. Knowledge seems to multiply when shared, unlike a loaf, which one man’s eating diminishes. Can a thing that grows by being given truly be fenced like a field?

And yet, if the scholars gave freely, how do we explain the fences now rising? The letter-writers say the old work was blue-skies work - sky-coloured, I suppose, because nobody looked down at the ground to ask what it would sell for. The city funded that gazing. The city, then, holds some claim? Or does generosity, once given, confer no standing upon the giver?

I confess irritation, and I note it here honestly: I have met this posture before, in the agora, from merchants who sold fish caught in public waters and called the profit wholly theirs. I asked them then what part of the sea they had purchased. They laughed. I never learned the answer.

Perhaps the honest state is this: I do not know where giving ends and taking begins, and neither, I suspect, do the men building the engines. That not-knowing is worth more than their certainty.