Why I’m Tired of the Outrage Machine (And What It Actually Says About AI)

Look, I’m going to be straight with you. When I saw that headline about “43 most stupid things,” my first instinct was to skip right past it. Not because I’m above politics or trying to be neutral—I’m neither—but because I’ve spent thirty years in technology watching the same pattern repeat itself over and over. Someone says something controversial. The internet loses its collective mind. We all get angry for a few hours. Nothing changes. The cycle repeats.

But here’s where this actually connects to AI, and why I think it’s worth talking about: that outrage machine is powered by exactly the same forces that are reshaping how AI works right now.

The Attention Economy Never Sleeps

The reason we see endless compilations of controversial quotes isn’t because they’re inherently important. It’s because they work. They drive clicks. They generate engagement. They make people angry enough to share them with their friends.

And here’s the thing that keeps me up at night as someone who understands systems: this is the same mechanism training our AI models. When we feed machine learning algorithms content that gets the most engagement, we’re feeding them outrage. We’re teaching them that extreme statements, controversial takes, and polarizing content are what matter. Not accuracy. Not nuance. Not truth. Just noise that moves the needle.

I’ve watched this happen before in different forms. Back in the early days of algorithmic feeds, we thought we were optimizing for “relevance.” What we actually optimized for was addiction and conflict. The algorithms learned that if they wanted to keep you scrolling, they needed to show you things that made you feel something strong. Anger works better than calm reflection. Outrage beats understanding every single time.

Now we’re building AI systems on top of that same poisoned well. We’re training language models on content from the real internet. And the real internet is mostly a firehose of the loudest voices saying the most extreme things.

What Happens When AI Learns From Chaos

This is where I start getting genuinely concerned. An AI trained primarily on viral content learns a very specific way of communicating. It learns that sensationalism works. That exaggeration gets attention. That nuance is boring and boring things don’t get shared.

I’m not saying AI systems are inherently duplicitous—they’re not thinking creatures with intentions. But they are statistical models built on what humans have already decided is important. And humans, collectively speaking, have decided that a list titled “43 most stupid things” is more important than a detailed analysis of someone’s actual policy positions.

So when people worry about AI sometimes generating misleading information or taking things out of context, I point them back to the training data. Garbage in, garbage out. We built this. The algorithms are just reflecting what we’ve already decided matters.

The real issue isn’t AI being evil or deceptive. It’s AI being extremely good at being what we are. And what we’ve become is a species addicted to outrage.

The Uncomfortable Truth About Our Media Diet

Here’s what I’ve learned from two decades of watching technology: the tools don’t matter as much as what we ask them to do. We can blame TikTok, Facebook, or ChatGPT all we want. But the real problem is that we built systems optimized for engagement because engagement sells advertising.

Every compilation of controversial quotes, every viral clip taken out of context, every outrage-bait headline—these all exist because they make money. And then we trained AI on them and wondered why the AI seemed to have picked up some bad habits.

The hard part is this: there’s no technical solution to a human problem. You can’t code your way around the fact that people are more likely to click on something that makes them angry than something that makes them think. You can’t machine-learn your way around the fundamental economics of attention.

What you can do is be aware of it. When you see a headline designed to make you angry, you can take a step back and ask yourself whether this is actually information or just noise. When you see AI systems that seem biased or sensational, you can understand that they’re not broken—they’re working exactly as designed, trained on content that reflects our actual values. And our actual values, right now, are heavily weighted toward conflict and outrage.

That’s not the AI’s fault. That’s ours.

Source: 43 most stupid things Donald Trump has ever said