AI Is Actually in a Losing Battle

With Amazon reportedly working on a home robot to be introduced by the end of  2018 and new books out like Human + Machine, the engine known as artificial intelligence (AI) appears to be gathering steam.

It would appear there is no stopping AI. Yet, yet at the same time I feel I must hold up a hand and say, “Wait a minute. This is not all it’s cracked up to be.”

AI’s Not So Intelligence

While this may come off as a bit of sour grapes, I lost out on a possible job two years ago because of Grammarly, an AI software tool that supposedly caught things I missed.

I tended to doubt the claim. So when I another member of one of my online editors groups posted this link about the shortcomings of Grammarly, I got a hearty chuckle from the story.

Headlined “Grammarly Fixed a Security Vulnerability, but It Still Can’t Fix Our Writing,” Slate author Jacob Brogan discussed a security researcher discovering a bug in Grammarly’s Chrome extension.

That threatened to let any website log in to Grammarly’s and access all your documents and other data.

Though Grammarly quickly closed the hole, Brogan went on to comment that nothing the company does could allay its underlying problem: “Its services just aren’t that good.”

He mentioned how another writer for Slate noted many of the “fixes” in Grammarly’s advertisements were either unhelpful or actively wrong.

Later, Brogan pointed out how Grammarly identified three potential problems in another article he published in Slate—and why each was wrong.

Word’s Not the Word

To be fair to Grammarly, Word’s spelling and grammar checker isn’t much better. In fact, I routinely ignore its suggestions and have made random notes about how some are flat out wrong.

Take, for example, Word telling me that neighbor should be spelled with an “our” on the end. Maybe in the United Kingdom, but certainly not in the United States (and my system is supposed to be set up for U.S. English).

It routinely advises that I place a comma after “so” when that isn’t so.

Sometimes a comma interrupts the flow of the thought or sentence, which is why I’m such a big believer in reading drafts out loud. Only then can you determine where, or whether, a comma should be placed in the text.

There’s more. With a book I’m presently editing, I noticed the author had used the word “causal” when he meant to use “casual.” Immediately, I assumed he must have relied on the Word grammar checker, which tries to substitute “causal” for “casual.”

Word also tells me to insert “its” for “it’s” when the latter is correct.

One time in place of “this feels” it offered “this feel” or “these feels.” At which point I thought, “Who in the world makes up these obviously wrong checkers?”

Matching Wits

I thought I was down on Word’s system, until I read the other article that Brogan linked to; it appeared earlier under the headline: “Microsoft Word’s Grammar and Style Tools Will Make Your Writing Worse.”

I won’t go into the obvious with a long recitation of Brogan’s critique; you can read it yourself.

My point here is that, no matter how fancy the bells and whistles, no matter how great the labor-saving device, no matter how awe-inspiring the technology, AI will never match the brain, mind, will, and emotions of the average person.

There’s a simple reason for that. Flawed humans are behind AI. The Creator of the universe designed people. I’m going with the Almighty on this one.