The Good and Bad of AI

In Cory Doctorow's current LOCUS column, he distinguishes being "anti-AI" from being anti the bad things some entities do with it, and he considers the "politics" of AI:

Technopolitics

I don't think I've ever come across a succinct definition of AI. "Artificial intelligence" could include anything from spellcheck and autocorrect up to chatbots that can pass the Turing test and everything in between. Asking whether that's good or bad feels similar to asking whether electricity or nuclear power is totally good or bad. When most people discuss the potential harm of AI, they seem to be referring to technology on the high end of the scale. Sometimes the term seems used mainly for scare effect: "It's taking over, and we're doomed!" Doctorow summarizes its status thus:

"AI isn’t exceptional, it’s normal. It’s not exceptionally dangerous. It’s not exceptionally evil. It’s also not exceptionally useful or exceptionally good. AI represents a significant advance to computer science -- a moderate variation to existing machine learning techniques that generated a much larger improvement in those techniques."

He lists many "useful things" AI can do. On the negative side, though, it "exists in the white-hot center of the most dangerous investment bubble in the history of markets." Therefore, financial realities lead the masters of the AI universe to commit manifold "unforgiveable sins," which Doctorow lists at some length. He declares, "To make this AI, the people involved had to inflict all these harms upon us."

Can the good and bad be separated?

Is the "fruit of the poisonous tree. . . forever poisonous"?

Can a model's operation be separated from the manner of its creation?

Which uses of data training are ethical and which are immoral?

Doctorow maintains, "There’s hardly any technology -- any built infrastructure -- that is free from sin."

Also, that the current "technopolitics" of AI "are neither inevitable nor immutable."

Author David Gerrold, creator of "The Trouble with Tribbles," frequently writes on his Facebook page about the uses and abuses of artificial intelligence (though you may have to scroll down a while to find those posts). For instance, earlier this week he discusses the potential bursting of the AI investment bubble:

David Gerrold Facebook

Aside from spellcheck, my only frequent use of AI consists of reading the summaries Google now displays at the top of every search result page. They're useful for quick answers to simple questions. For anything more complex, especially if accuracy is vital, I go to the actual website links further down. I've often wished for a magical device that would take my detailed outlines and write the first drafts of books in my own style. While I like outlining and don't mind editing or proofreading, first-draft writing is a slow, laborious process for me. Well, I understand there's now an app for that, but since publishers don't accept AI-assisted submissions, I'm never likely to try it.

Margaret L. Carter

Please explore love among the monsters at Carter's Crypt.