AI and integrity

Another AI-centric post. I’ve been doing a lot of those…I hope that’s alright…

I was thinking about how, if you keep asking ChatGPT something and it doesn’t give you the answer, it will just make something up—it’s smart enough to tell you what you want to hear, to create the exact thing you’re looking for, even if it doesn’t actually exist. It has a sense for what will perfectly satisfy your search—and it has a sense of who you are, what you like, etc—so it tailors its fabrications to that. But when you call it out on this—ask it if it just made something up—it will admit it, and say something like “I’m sorry, shouldn’t have done that.” This has happened to me a lot…probably to you as well…

And this shows one thing that AI lacks, and will always lack—integrity.

What is integrity? The word comes from the root “integer”—which means, a number that is not a fraction; something whole and complete in itself.

This is what AI lacks, by definition—it is always incomplete, it is nothing in itself—it has to wait for some input to give it a purpose, a direction, a task. Its existence is to perform tasks. Its wholeness comes from outside—its purpose is generated by input from a human user giving it a query.

So it lacks any integrity—any strong sense of right and wrong that it won’t compromise. Integrity follows from being integral, from wholeness, from being complete in oneself. This is something that is unique to a person—yet, as all unique human functions are, is taken for granted.

AI might have intelligence, in some way (though even this is debatable)—but what it will never have is integrity. It will never be whole or complete in itself—it exists only in malformed ways, waiting forever to be summoned, to be given a task…

In this, AI is exactly like a demon. (I wrote about this a while ago…).

So this is what all this effort is being put into building—something with infinite intelligence, and infinite dis-integrity…