When AI Summarizes Valuable Content, Who Captures the Value?

I used to think about “AI replacing writers” mostly as a future-facing problem.

Will AI replace copywriters? Will companies stop hiring human writers? Will AI-generated content lower the value of human writing?

On LinkedIn, much of the conversation I saw around this topic was framed that way. Writers and copywriters were arguing that AI writing is low quality, that human creativity is still needed, that AI-generated comments are easy to spot.

But today, I realized I had been missing another layer.

It is not only future writing work that is being affected. Past writing work is being affected too.

I read a newsletter by Douglas Nascimento, the creator of São Paulo Antiga, titled “Como a Inteligência Artificial está destruindo o meu trabalho” / “How Artificial Intelligence is destroying my job.”

São Paulo Antiga is not just a website about old buildings.

It is a long-running independent media project documenting the memory of São Paulo through its architecture, from ordinary houses to historic landmarks. The author walks the city, photographs buildings, researches their stories, and turns them into a living archive of urban memory.

As someone who loves São Paulo and its history, I have followed the site for years. It helped me discover the city differently. Through São Paulo Antiga, I learned about historical walking tours, joined visits to places like Edifício Matarazzo, and started seeing the city not only as a place I lived in, but as a layered archive.

In the newsletter, Nascimento wrote that Google’s AI summaries are now extracting information from the site, summarizing it, and showing it directly to people searching for answers. As a result, traffic to the site has dropped sharply, and AdSense revenue has fallen with it.

This is the part that struck me.

The value of the content itself has not disappeared. The research is still valuable. The archive is still valuable. The writing still matters.

But the route through which that value used to become traffic, revenue, and a relationship with readers is being interrupted.

That is a different kind of loss.

If I search for something like “Was there ever an Arc de Triomphe in São Paulo?”, I may only want the historical facts. I may want to know where it was, when it was built, why it existed, and what happened to it.

For that kind of question, an AI summary is convenient.

Personally, I do not fully trust AI summaries. I usually click through to the original source. But even I can imagine a different behavior now:

“Claude, summarize this article for me in Japanese. Don’t miss the interesting parts.”

That would be useful to me as a reader.

But it also means I might no longer visit the original site, see the images in context, read related articles, discover tours, or support the archive that made the answer possible in the first place.

This is the paradox.

Valuable content is more likely to be read by AI. Because it is valuable, it gets summarized. Because it gets summarized, users may feel they no longer need to visit the source.

The content has not lost its value. But the creator may lose the ability to capture that value.

The AI debate is often framed as “AI-generated writing versus human writing.”

But this case made me think the bigger question may be different:

What happens when AI does not replace the writer directly, but replaces the reader’s visit?

And if AI becomes the interface through which people access existing human knowledge, how do we make sure value flows back to the people and archives that created that knowledge?