My Views on AI writing


I am seeing a lot of current competitions that are banning the use of AI completely, and I have mixed feelings about this.

On one hand, I see what they are trying to do. The rise of AI-generated slop, not to mention the disturbing level of AI use even on forums like Reddit and Quora, in emails and in social media posts, is annoying at best and a serious concern at worst. Scammers are more sophisticated with their phishing. I miss the Nigerian princes. And the American soldier in Iraq who found a ton of gold. Some of those tales were actually pretty good.

Fake news, which was always a concern, now requires even lower effort to generate, and the content is really mind-numbingly stupid.

In this environment, it makes sense to carve out a space for ourselves and declare this off-limits to AI. To insist that writing is exclusively for humans.A place where writing is supposed to come from sweat, misery, caffeine, and years of craft. Soulless machines have to wait by the door.

Noble. Slightly medieval. utterly doomed to fail.

Here’s the problem.

AI is here to stay.

The genie escaped the bottle, set up a premium subscription service, and is now offering you personalized recommendations. 

Expecting people to honor an AI ban because they promised they wouldn’t cheat is… adorable? These organizers seem to have a touching faith in human nature that’s wildly inconsistent with, say, the entire history of human nature.

But let's set that aside for a moment. let's understand the nuance behind what constitutes AI use. 

If you’ve used Grammarly, is it okay? 

What if you’ve used AI to clean up punctuation, grammar, and typos? Does that count?

What if you've asked it to help you with a couple of similes or metaphors? Is that considered a problem?

LLMs are not, at this point in time, sophisticated enough that they can write full-fledged stories anyway. Most of the time, they produce something that reads like it was written by a competent intern who has never actually experienced emotions. You still have to do the real work.

The actual danger isn’t that AI writes masterpieces.

It’s that it gets just good enough that we can’t reliably tell the difference anymore. With every successive iteration, our ability to distinguish between AI-generated content and genuine human writing becomes blurrier.

Mark Lawrence ran an experiment asking people to separate AI-written stories from human ones. Many of us — myself included — couldn’t do it with any confidence. Turns out our finely tuned artistic instincts are about as reliable as a coin toss.

A blanket ban doesn’t solve anything. It just encourages people to get sneakier.

If you really want to limit the use of AI, make two categories — one for AI writing and one for non-AI writing. You’ll still have a few Charlies smuggling their work into the wrong bucket pretending to be human. Humanity has always produced Charlies. That’s tradition.

But you have another bucket full of AI slop and maybe an occasional diamond in the metaphorical trough.

Feed both piles into detection models. Train better classifiers. Use each contest to improve the next one. Crowdsource the problem-solving.

Fight fire with fire.

Or more accurately: fight statistics with slightly more sophisticated statistics.

It’s not perfect. It won’t be clean. But it’s a lot more practical than pretending a strongly worded rule will keep technology at bay.

Because history suggests that when humans meet a new tool, we don’t abstain.

We give it a gentle poke. Then a prod. 

We might start praying to it and sacrifice a few goats in it's name, if we dont understand it. That's also depressingly normal for humans.

 And if it makes life moderately easier, we build our whole civilization around it.

You know.

The way evolution intended.

Comments

Popular posts from this blog

An Article by the Asian Review

I Spent $1300 on Amazon Ads— Here’s what I learned

The pursuit of the perfect...desert?