My parents were both tier-one journalists in Italy. My grandfather was editor-in-chief of the country’s most important newspaper. And I started my own career in journalism, learning how to find reliable sources, structure a story, and ask the question that matters most: What’s the point?
Lately, as the advertising and media industries reckon with the rise of generative AI, that question keeps coming back to me. Everyone is talking about how AI is disrupting publishing. But too few are asking what the point of AI is without publishing.
This is the angle we too often miss: AI companies rely on publishers.
And if we don’t find a sustainable value exchange soon, we risk degrading the very thing that made AI useful in the first place: a rich, open, diverse supply of high-quality human content.
Cloudflare’s “Pay-Per-Crawl” Model Is a Warning Sign
Last summer, Cloudflare made waves with its “pay-per-crawl” proposal, a new model that would allow publishers to charge AI companies for access to their content at the infrastructure level. Much of the conversation focused on copyright and data licensing.
But the bigger story isn’t legal. It’s economic.
If creators no longer benefit financially or reputationally from producing content, they’ll stop. And when that happens, AI loses, too.
The quality of training data erodes. LLMs start learning from content that’s increasingly synthetic, repetitive, and low-value. A vicious cycle begins, one where the open web is hollowed out, and generative tools start feeding on themselves.
This is not a hypothetical. Google rolled out AI summaries in its Discover feed. Traffic to publishers is dropping. Monetization is harder than ever. And many publishers, especially smaller ones, are asking whether it’s worth continuing at all.
You Can’t Automate What You’ve Scraped Into Extinction
We’re living through what I’d call an AI land grab. Every platform is racing to integrate generative capabilities to create content, plan campaigns, and build audiences. In some cases, the results are genuinely useful. In others, they’re hype.
But across the board, the strategy is extractive. AI companies are pulling value from the open web — the product of decades of human reporting, editing, and analysis — and giving little back in return.
I’ve spent my recent years in ad tech, but I still think like a journalist. And I can tell you this: content doesn’t make itself. It takes time, people, and infrastructure.
If the economics no longer work for media companies, if original reporting can’t be monetized, then publishers will either cut corners or cut content entirely.
The end result would be that the signal-to-noise ratio collapses. AI systems trained on this degraded content would start producing worse outputs. We would all lose: readers, advertisers, tech platforms, and yes, the AI companies themselves.
This Is Everyone’s Problem, Including Advertisers
For marketers, this might sound like someone else’s issue. But it’s not.
The success of any AI-enhanced marketing strategy, whether it’s media planning, content generation, or audience segmentation, depends on the quality of the underlying information. If the web becomes less trustworthy, less diverse, or less current, your AI tools become less useful, too.
Worse, if publishers disappear, the internet becomes less navigable. Brands lose context. Media placements get riskier. The cost of customer acquisition rises.
In short: advertisers need a healthy publishing ecosystem just as much as AI platforms do.
Ad Tech Has a Role to Play
As someone who now works on the infrastructure side of advertising, I see the connection clearly. Ad tech is built on the open web. And for too long, it’s benefited from that openness without helping enough to preserve it.
That needs to change.
Publishers can’t rely on CPMs alone to fund high-quality journalism. Nor should they have to fight off bots, scrapers, and spam networks while figuring out how to get paid for what they produce.
Whether through curated supply paths, smarter measurement, or partnerships with AI platforms, the advertising industry can and should help rebuild the economic foundations of original content.
We’ve already seen hints of this. Some platforms are reinvesting infrastructure savings to reduce take rates and pass more value to publishers. Others are experimenting with flat-fee models or performance-based guarantees. These are steps in the right direction.
What Comes Next
I’m not naive. I know that media companies and AI platforms won’t suddenly become best friends. There will be friction. There will be lawsuits. There will be bad actors.
But there’s also an opportunity to get this right, to build a system where content creators are compensated, platforms are transparent, and AI systems are accountable to the quality of their inputs.
That starts by remembering one thing: you can’t generate insight from a knowledge base that no longer exists.
AI doesn’t just need data. It needs journalism. It needs publishers. And it needs a sustainable ecosystem to keep the information flowing.
If we destroy the source, we destroy the system.
