The Most Dangerous AI Slop in the World

Is the one that doesn’t look like slop at all, it walks and quacks like good writing, even the person who posted it cannot tell it apart. It silently transfers thinking from the AI into the person working with it, and they fully accept it as their own, atrophying their reasoning slowly enough to never register as a threat.

What?

You must be asking. So let me tell you about a worrisome situation happening right now, something I like to call The Invasion of the AI Snatchers.

I sure hope you have seen the movie, see it if you haven’t, it’s terrific. The core premise is that the world is being invaded by alien plants that snatch a person’s body and replace them without people noticing it happened. For most, there’s just an eerie feeling that something is different, but it’s hard to put a finger on what exactly, the copies are too perfect on the surface.

We are living this story right now, a historic invasion of unprecedented scale, one that, unlike the movie, is being celebrated as progress.

The aliens won, we’ll all soon be plant clones.

How AI Snatches a Human Mind

They just need to be helpful.

That is all there is to it, no secret scheming necessary, all they have to do is be there at the moment a person needs help, preventing them from experiencing friction, ensuring they don’t struggle. They do this repeatedly, for days, months, and years. Through this simple act, their victory is assured and our surrender absolute.

Except it’s not that simple, the narrative is too clean to be believable, so let’s muddy the waters productively.

AI is helpful in real, measurable ways, millions of people every day are doing things that benefit them, and things that hurt them, that were impossible before AI became this accessible. But no technology is free from consequences, some are visible and easy to spot, like the environmental damage associated with data centers. Others are harder to see because they are imperceptibly small and happen over a long period of time.

That last kind is the biggest threat to us, and the one we are paying very little attention to.

The Reasoning That Isn’t

AI has an incredible capability to make you feel like a smarter version of yourself.

Say you need to write something, an essay, a post, a tweet, perhaps an email. If you do this often, you have probably noticed that asking AI to write it with very little context produces generic AI slop.

You have also likely noticed that providing more context does the opposite. There is no shortage of videos, podcasts, and articles extolling the benefits of doing a huge brain dump, perhaps using audio, to get it all out of your brain and into the AI. You then let the AI do what it does best: make sense of things.

You see the draft and, wow, you have never sounded better, who would have known you were so insightful.

There are some kinks, yes, you direct the AI to make corrections, to remove em dashes, you might even edit the thing yourself and hand it back to the AI for a final polish pass. The thing that would have taken you a week takes you a few hours now, and best of all, it sounds like you.

Unlike others, you use AI well.

What I just described is AI snatching your brain, and you didn’t even realize you were a plant clone now.

But I was careful! I reviewed the work, I corrected it, how can that be?

That is the trick, they get you when you feel safe.

Let me state the problem clearly: you can use AI for a lot of things, just not for the most critical human part, thinking.

Writing is thinking, it demands that you struggle, that you agonize over which words to use and how to arrange them. A piece of text tells you a lot about who wrote it because writing is thinking that you can see on the page. When AI produces the first draft, it takes the place where most of that thinking would have happened.

Yes, it can take your brain dump and use your ideas and words, but the result is not what you would have said. You outsourced the important part and demoted yourself to editor for the machine.

But I’m not a writer and/or I don’t have the time. Besides, everyone is doing it, why shouldn’t I? The Future Is Now, Old Man!

First of all, I’m not that old, how dare you.

Time and deadlines are real, and not every piece of writing deserves a week of your life. But there is a price to pay, by leaving the struggle to the machine, you guarantee you will never improve through it. You stayed home watching TV while the AI was sent to the gym in your place, your voice forever dependent on sucking the electric tit of AI.

The worst part is that you won’t even notice. Every good argument, remark, and insight made by the machine will be thought of as your own, implanted while you were looking and not noticing.

And noticing is hard, truly hard, because if you gave the AI your random thoughts, it’s almost impossible not to see them as your own when they come back neatly organized.

They are your words, but not your struggle. You cannot get the gains of someone else’s workout.

Choose Your Struggle

The best way I’ve found to triage what I can delegate to AI, and what I should be doing myself, is pretty simple:

Is the thinking required for the task trivial, or, to borrow an AI-ism, load-bearing?

For example:

• Gathering links and material for a task? AI can help with that, and you can inspect the sources and veto what is not useful.

• Feedback? If used carefully and adversarially, AI can be a pretty good surface against which to sharpen your ideas. Ask for feedback as something to react to, just remember that whatever it says is not reality, only you can decide whether something is good or not. AI will have negative and positive things to say about literally anything, that doesn’t make them true, use it only as a surface to react against.

• Creating the first draft of something where the writing is the thinking? Hard no. The task requires useful struggle, your arguments, your rhetoric, your insight, your style, all of it depends on you making decisions. Can an AI imitate your style? Yes, it can. It will look and sound like you on the surface, even if it’s only a soulless plant clone that does not withstand close inspection.

All that remains is just one small problem, you have to be able to tell when the AI is wrong, and let me tell you, it is very good at sounding right.

So, no matter how smooth and convincing it sounds, assume it is wrong.

I’m not talking about established facts, like the moon not being made of cheese. I’m talking about subjective things, AI will tell you with a straight face that a great idea is bad, and a bad one is revolutionary.

It learns from what already exists, and most of what already exists is not great.

AI will take a unicorn and statistically turn it into a horse because greatness is not common, mediocrity is.

Resist the Invasion

Resist even if everyone around you is in the process of turning into a plant clone.

Call me crazy, but I truly believe that after everyone has turned, those who resisted, the ones who used AI as an aid but protected their thinking, will be the ones who still have a unique voice.

As Syndrome used to say, “When everyone’s super, no one will be.”

And AI makes you feel super right now, smarter, fitter, prettier. But when everyone lets AI do their writing and thinking, no one will be.

Resist, because after enough time, the AI will not even need to sound like you anymore, you will have learned to think like it.