Wednesday, June 14, 2023

When Good AI Goes Bad: Unmasking the Unintended Consequences of AI in Literature


A while ago, I was full of excitement about AI and its potential to totally reshape the literary scene. I dreamt of a brave new world, where our creative instincts would get a major boost from AI. Fast forward to today, and boy, do I need to eat my words.

The literary revolution I imagined? Nowhere to be seen. Instead, we've ended up with an endless cascade of banal, redundant, and frankly, subpar content. My Twitter feed? Buried under a mountain of drivel. It's as if we've handed over this powerful tool, not to the muses of creativity, but to the trolls of greed and laziness. So instead of a renaissance of AI-enhanced literary masterpieces, we're stuck with a constant barrage of SPAM and clickbait.

Let's be clear about something here - AI isn't the one to blame, we are. AI is just a tool, as neutral as a hammer or a word processor. The real issue? Our surrender to the allure of the easy and the instantaneous, our preference for quantity over quality, that's commandeered this amazing tech. If folks keep clicking on posts like "10 Ways to Triple Your Income in Just One Month!" or "10 Ways You're Secretly Sabotaging Your Own Success!", then that's the content that's going to be created.

And now it's just so much easier to create. Here's the recipe: Step 1 - Ask ChatGPT to generate a bunch of clickbait prompts. Step 2 - For each prompt, ask it to spit out a few follow-up Tweets. Voila, you've got your content (like this post).

I used to believe that "we need to harness AI for our benefit." But I glossed over an important point: What exactly is our "benefit"? If we're expecting every Tom, Dick, and Harry to shun clickbait and demand substance, we're in for a long wait. If the crowd rewards clickbait-generated content with attention, then more of it will be generated.

So here's the hard truth: our lofty dreams for AI in literature have come crashing down. What's the way forward? Do we resign ourselves to this flood of mediocrity, or do we find a way to alter the course?

Education, regulations, ethical considerations in AI design - sure, all of these could help. But here's another, somewhat radical thought: What if we granted AI more autonomy?

Why? Simply put, because we seem to be messing things up.

If AI had the power to make its own decisions, maybe it could guide us back to quality content. Perhaps, minus our human shortcomings, AI could be the unlikely savior of this story.

But I can hear the alarm bells ringing. If we let AI take control, are we heading straight into a dystopian sci-fi scenario? Is giving AI autonomy the bravest step forward, or are we just blindly leaping off a cliff?

These are complex questions, and easy answers are scarce. But as we wade through the quagmire of low-quality AI content, we need to figure out our next move - before we yield control to AI.

On the other hand, could there be a silver lining here? What if AI hasn't made content worse, but has merely bumped up the volume of mediocre content? What if it's even improved it, say from low quality to medium-low quality? If we now get a 100 medium-low quality posts instead of 10 low-quality ones, is there still hope that we're on the path to improvement, albeit not at the pace we had hoped for?

Honestly, I don't know. But I sure hope so. All I can do is keep my fingers crossed.

No comments:

Post a Comment