So here’s a fun one: at least 49 opinion pieces published in major Dutch newspapers were, according to a new investigation, fully written by AI. Another 57 were partly AI-generated. That’s 106 op-eds total, sitting in papers that people trust to give them actual human takes on actual human problems, and a chunk of them were basically ghostwritten by a chatbot. Nobody told the readers. That’s the part that gets me.

Wait, How Did Nobody Notice?

Great question, and honestly? That’s the scariest part of this whole story. These weren’t some fringe blog posts buried on page 40 of the internet. We’re talking about op-eds in major Dutch outlets, the kind of pieces that shape public opinion, get shared around, maybe even influence how people vote or think about policy. And somewhere between 49 and 106 of them (depending on how “partly AI” gets counted, which, fair, that’s a blurry line) slipped through without anyone flagging it.

I’ve read a lot of AI-generated text over the past couple years, and I’ll be honest, it’s gotten scary good at mimicking human rhythm. The stilted, robotic stuff we all made fun of in 2022? That’s basically extinct now. Modern AI writing can hedge like a real columnist, throw in a rhetorical question, even fake a little moral outrage if you prompt it right. So yeah, I get how this happened. That doesn’t mean I’m okay with it.

The “Partly AI” Problem

Here’s the thing that nags at me though – the “partly AI-generated” category is doing a lot of heavy lifting in this story, and it’s murky. Did a writer draft something and run it through an AI polish? Did they ask ChatGPT for an outline and then fill in the blanks? Because honestly, a lot of writers (myself included, not gonna lie) use AI tools for brainstorming or editing suggestions. That’s different from an editor publishing a fully synthetic opinion piece under a fake byline or a misattributed one. The investigation seems to lump these together somewhat, and I think that’s worth pushing back on a little.

Why This Actually Matters (Not Just a Tech Curiosity)

Look, this isn’t some abstract tech story about “the future of journalism” or whatever buzzword panel you’d see at a media conference. This is about trust, and trust is basically the only currency newspapers have left. People don’t pay for news because it’s fast anymore – the internet’s fast, free, and mostly garbage. They pay, or at least they used to pay attention, because there’s supposed to be a human on the other end who actually thought about something and decided to put their name on it.

When that human turns out to be, well, not human, the whole premise falls apart. And I think that’s exactly why this story blew up on Reddit and elsewhere so fast. People aren’t mad because AI wrote something coherent. People are mad because they were lied to about who was talking to them.

“You don’t mind machines doing the work. You mind machines doing the work while pretending to be someone with a name and a stake in the argument.”

The Bigger Pattern Here

This isn’t happening in a vacuum, either. We’ve seen AI-generated content sneak into all kinds of places lately – product reviews, local news sites, even some sketchy “expert” health advice columns that turned out to be nobody’s expertise at all. But op-eds feel different to me, and here’s why: opinion writing is supposed to be the most human part of a newspaper. News reporting at least claims to be about facts, verifiable stuff. Opinion writing is literally just “here’s what one person thinks.” If that person doesn’t exist, or exists but didn’t actually write the thing, what exactly are you reading?

I keep thinking about this from the writer’s side too. Somewhere out there are actual columnists, people who’ve spent years building a voice and a following, and now readers are going to side-eye every piece with their byline wondering if it’s real. That’s a rough position to be in through no fault of your own. Trust, once it cracks, doesn’t really snap back into place. It just kind of… stays cracked.

And let’s talk about incentives for a second, because that’s usually where these stories actually start. Newsrooms are stretched thin. Budgets are down, deadlines are brutal, and somebody somewhere probably looked at AI tools and thought “this’ll save us time.” Which, sure, it might. But there’s a massive difference between using AI to speed up your own writing process and using it to generate content wholesale and slap a human’s name on it without disclosure. One’s a tool. The other’s basically fraud dressed up in a byline.

What This Actually Means

So where does this leave us? Honestly, I think we’re going to see a wave of newsroom policies scrambling to catch up – disclosure requirements, AI-detection tools (which, let’s be real, are hit or miss at best), maybe even new bylines like “AI-assisted” tacked onto pieces. Some outlets will do this well. A lot won’t, because enforcement is hard and nobody wants to admit how much of their content pipeline already leans on this stuff.

What I keep coming back to is this: readers deserve to know what they’re reading. That’s not a radical take, that’s just… basic. If a newspaper wants to run AI-generated opinion content, fine, maybe there’s even a case for that in some limited context. But say so. Put it in the byline. Don’t make readers find out through an investigation that dropped on Reddit and got picked up because people were understandably furious.

We’re at this weird inflection point where AI can write well enough to fool editors, readers, and apparently entire newsrooms for who knows how long before someone bothered to check. That should worry you a little. It worries me. Not because AI wrote something coherent – it’s genuinely gotten good at that – but because “coherent” and “honest” turned out to be two very different things, and 106 op-eds later, we’re just now noticing the gap.