AI Is Swallowing the Free Internet

By Morgan Ramsey June 13, 2025

This isn’t a new story.  A massive technology breakthrough comes along that mesmerizes the world, and there’s a global arms race to figure out the fastest way to monetize the new invention. 

Everyone is too busy figuring out how to get rich off the new tech that they forget to think about the long run, real world impact of the new business model they worked so hard to figure out.

Everyone sees the potential power of this new tool, but the focus isn’t purely on using this new innovation as a source to make us all smarter and more connected. The focus is on gaining market share and monetizing traffic.  And if you care about independent writing, journalism, or just the idea of a halfway-intelligent internet, you should start paying attention.

A Wall Street Journal piece this week showed what anyone running a media site already knows in their gut: search traffic is tanking. The stats are staggering.  Some outlets are seeing 40-50% drops in search-driven clicks, and news rooms are responding by laying off employees.  Just last week Business Insider announced it was laying off 20% of its staff to focus more on, what else, AI. Whole revenue streams are drying up overnight and news sites are trying to figure out ways to keep the lights on.

Why? Because users are no longer getting pushed toward the open web. They’re getting slick little AI-generated summaries and answers — straight from the search engine, no click required.

In other words, the internet is cannibalizing itself.


Game over for free content model

For the last 15 years, the open web has run on two basic monetization models - ad supported free content, and subscription based paywalled content.  There hasn’t been any middle ground.

This was always an unstable system, and as it became more unstable a lot of news sources responded by focusing on outrage over depth and truth. But now that system is breaking down completely.

Here’s where it gets tricky. Moving to a full subscription model — a hard paywall — works for maybe 1% of writers and publishers. The New York Times can do it. A few high-profile Substacks can do it. Most can’t.

The problem is that the middle option doesn’t exist.

Right now, there are basically two choices.  Give all content away and hope the ads cover it (an increasingly difficult proposition).  Or lock it all up and hope you can convince people to subscribe.

Unfortunately there’s a limit to how many subscriptions anyone can reasonably maintain.  I love substack but I already subscribe to two major publications and about 4 Substacks.  I’m personally maxed out.

There’s no easy, frictionless way for readers to just pay for the content they care about, from writers they trust, about topics that matter to them.

Why can’t I throw two bucks at a great investigative story my friend shared without signing up for a monthly plan? Why can’t good writers get paid directly, without having to funnel every reader through a subscription funnel?

That’s what we’re hoping we can fix with LedeWire.

If we can give readers a simple way to compensate the creators whose work resonates — no commitment, no hassle — I think the outcome can actually be better than what we’ve all grown accustomed to over the past 20 years.


The AI trap

One more thing to chew on. AI may be great at stitching together sentences, but it can’t do original reporting. It can’t talk to sources. It can’t chase a lead down a back alley at 2AM. What it can do is remix the reporting and writing of actual human beings — the ones currently getting squeezed out of the ecosystem.

But journalists need to make a living.  And everyone that cares about being informed, that wants to really understand the topics that are moving our world, needs to care about this. 

AI can’t survive without a clean data source. If all it has to feed on is rage posts on twitter and reddit discussion forums, the answers it gives us will be garbage — and that garbage will compound, as the next generation of AI models is trained on an internet already stripped of quality reporting."

If AI companies want to keep training on real journalism, they’re going to have to start compensating the people doing that work. Full stop.  I think Ledewire can also be a conduit to make this happen and support journalists when AI models access their work to produce answers.


Why this matters

I’m not naive about this topic.  I know it always sounds good to say we want to be better informed.   But when push comes to shove it’s always easier to check your instagram feed over reading a long form investigative piece.  But there is still a appetite, and I would argue a dire need, for high quality reporting and investigative work.  At its best, the open internet can be a place where independent voices can break through, where great reporting can spread on merit alone, and where readers have a way to find free and unbiased voices that challenge and inform our opinions.  (And yes there’s even a place for the occasional cat meme or fail army video) 

We’re at risk of losing the promise of the open internet era, where everyone with access to a phone or a computer could have access to the best and brightest voices in the world.   

We started LedeWire because we believe there’s still a way to get this right. A way to fairly pay writers, creators, journalists that are doing the hard work that AI can’t do. A way to give readers more choice and better access to real information. A way to ensure the AI future can help us all understand the world better and access knowledge in a way that was unthinkable a few decades ago.

We’re not there yet, but we need to try.