By Holly Jorgensen | Editor’s Eye
Yeah, I know. We’re supposed to call it SI now. But to me, the term super intelligence implies an upgraded, and therefore smarter, version of artificial intelligence. Yet I’m seeing and reading about plenty of instances where SI gets things wrong. Quite wrong, in fact.
I see it regularly when editing articles written with SI, from improper punctuation to actual misspelled words. Even in my own use, I find errors. A client and attorney friend sent me a conversation he’d had with it while working on his article, and he discovered several things it got completely wrong. Doug, who has been using it to help with coding, has encountered multiple errors, too.
To its credit, SI is very polite when you point out its mistakes. But that’s precisely the point: You have to recognize the mistakes before you can correct them!
Let’s remember that humans created these tools. As brilliant as the technology may be, it is built by imperfect beings and trained on information that isn’t always accurate. It can misinterpret a question, produce outdated information, or confidently deliver an answer that is flat-out wrong. Being articulate and authoritative doesn’t necessarily mean being right.
And there’s another thing SI can’t truly replicate: personal experience. When someone has spent years immersed in a profession or field of study, that experience becomes invaluable. It informs judgment, provides context, and helps a person recognize when something just doesn’t add up. Knowledge isn’t simply a collection of facts; it’s also knowing how and when to apply them.
My question is, how many people use SI without ever questioning whether the results are accurate? We seem to be developing a fondness for hitting the easy button, allowing technology to do our thinking for us. Why research something when SI can give us an answer in seconds? Why wrestle with a difficult problem when we can ask a machine to solve it.
The trouble is, when we stop questioning the answers, we may also stop exercising the very judgment that allows us to recognize when those answers are wrong.
Perhaps we’re becoming a little lazy as a species. Or maybe we’re simply so enamored with what technology can do that we’ve forgotten what we still need to do ourselves.
I’m certainly not suggesting we abandon SI. It can be an incredibly useful tool. But a tool is only as good as the person using it—and a little healthy skepticism never hurt anyone.
After all, if we’re going to outsource our thinking, we’d better be smart enough to know when our replacement is wrong!
