Some believe AI will fundamentally reshape how we read. Instead of starting with a book or article, people might begin with a summary generated by AI. From there, they’ll explore the topics that interest them most, digging deeper through questions and prompts. Whether they ever engage with the original material might not matter.

Picture this: You’re curious about a non-fiction book. Rather than reading it cover to cover, you ask AI for an overview. It gives you the central arguments. You ask for a summary of key topics. As something catches your attention, you dive in further—requesting quotes, examples, counterpoints. Eventually, your questions outpace the book, and AI branches into other sources to continue the conversation.

So, who read the book? You didn’t, not directly. But you still got what you wanted through an AI middleman.

If this becomes the norm, authors might stop writing for human readers. Instead, they’ll write for AI—crafting content that helps it extract and deliver ideas clearly to the end user.

And it’s not just books. People are already using AI to learn instead of going to the original sources. That means if a company or individual wants to shape how they’re perceived, they’ll need to make sure AI knows them well.

Take buying decisions. Why would I schedule a call with a sales rep when AI can already answer my questions about a competitor?

I’m talking about the unglamorous but crucial stuff:

  • Can this product support my company’s specific business model?
  • Does it track the specific metrics my business cares about?
  • What are the workarounds?
  • How else could it help me?
  • What are the payment terms?
  • Is there a contract?

If AI can’t answer those questions about an offering, but it can for their competitor, the first company is behind. The companies that train AI best will win.

That means publishing information not just for people, but for language models. Even content you don’t expect humans to read directly can be useful if it helps AI serve the right answers to the right people.

There are risks. What if AI learns something outdated about you? How do you correct it? Right now, there’s no great answer.

But one thing is clear: more and more, people are getting their information from AI, not original sources. The way we read is already changing.