Tuesday, July 9, 2024

AI and the "either/or" fallacy


Credit to the artist of the cartoon, whose name is Beancius if I’m reading it correctly.

“There are only two kinds of people when it comes to AI: Those who see it as the future, and Luddites.”

Someone in a college administration said that. They’re a “young gun,” and I remember being like them, excited about every single new technology arriving on the scene and what it can do with “data.” I still can be that way, and I can certainly understand the excitement for artificial intelligence.

But I also know an “either/or” logical fallacy when I see one. Closer scrutiny of the situation is more apt to bring someone to somewhere in the muddy middle of that kid’s assertion…as is so often the case.

See that cartoon? Upon close reading, I’m sure you’ll find it quite clever. It also encapsulates my greatest concern regarding AI, namely that we will come to entirely outsource our thinking. You ask for what you want and it gives it to you, and you take it without thinking. I see a similar behavior all the time with students and others who do a Google search on a given subject, grab the first three entries that bubble to the top, and think they have all they need to know. It will be tough to persuade me people won’t carry out the same behavior with AI only at scale.

Kurzweil’s new book, according reviews in the Times and The Economist anyway, utters the common refrain that AI will free us from mundane tasks. Cool. He goes on to say that millions of workers will be “liberated.”

That’s an odd, Orwellian term for “unemployment.”

Because he also predicts that in five years, AI will be able to write whole books and create art indistinguishable from human work. If so, just what will my family and I to do with all of my newfound “liberation?” Said “liberation” would, in fact, be a loss of self. Libertarians might argue, “Suck it up, buttercup. Market success means changing what you do.” Well, changing what I do is pretty much the same thing as changing who I am.

Ain’t so easy. I don’t think it will be easy for a lot of other people, either, even coders.

There are amazing things in the offing with AI. Medical treatments. Scientific discoveries. Anything that once required humans to take painstaking efforts to see patterns in data can be done in comparatively brief amounts of time. Good stuff. I just don’t think it makes one a “Luddite” to want to take a critical eye to what’s happening, and maybe pump the brakes for a moment.

At least we won’t ever have to worry about an AI being categorized as “alive,” or so a biologist argues from a mechanistic perspective. (see comments)

But hey. We’re getting an AI-generated Al Michaels for the upcoming Olympics. So we got that going for us.

Which is…nice?



Follow me on Twitter: @Jntweets