Meredith Kopit Levien, the CEO of the New York Times, was interviewed by Axios’ Sara Fischer at Cannes and explained how the paper survives. She’s really good at this. She went through the products people seek out by name, a wide free layer, the bundle, a family control structure that lets them think in generations, and a growing video investment.
They’re not really separate advantages, though. They describe the same, single advantage.
I’ve spent my career watching the attention economy emerge from the inside, building strategy around where an audience actually comes from. And I’ve watched this ground shift twice. Facebook gradually cut off traffic to publishers and websites after 2016. Then a lot of newsrooms laid people off chasing video numbers that Facebook later admitted were inflated.
Now Google is pulling back search referrals, with its new AI results answering your questions directly instead of sending people to the sites that invested in the work.
The thing with the Times is that a large number of people come to it directly, without a newsfeed or search result in between. They type the address into the browser or go to their bookmarks, they open the app out of habit, they go straight to Wordle when they wake up. The free games, the bundle and video are all real, but none of them is rare, and some of it is overstated. Plenty of companies give away a free layer or sell a bundle, but almost none have a direct relationship with millions of people who show up without being sent.
This direct relationship is exactly what the AI companies are designed to capture. And that’s where this turns into a story about power.
OpenAI, Microsoft, Anthropic and Perplexity built systems that sit between a reader and the newsroom, so the reader gets what they came for and never has to make the trip. Perplexity and others answer the question without a click. You’ve seen how Google does the same now. Media companies still pay for the reporting, but the AI model racks up the impressions.
All of this happened without any kind of public process. There wasn’t a hearing or industry input. These companies looted the archives of the entire press (and others) and let everyone find out after the fact. The only place left to fight back is the courtroom, and only publishers with real money could afford to act.
Levien and the Times have spent $20M and 2.5 years in court over the scraping. These AI companies didn’t take all that work just to scrape it. They’re building something that answers your question for you, so you never have to go to the Times. The AI product will be the place people go, and they want people to trust them the way they trust the Times. But without putting in the 175 years.
Look, the Times will probably be fine and continue to thrive. It has the direct relationship, the 175 years, the lawyers. It’s the “paper of record.” That’s not the part to worry about. We have to look out for the middle players, the niche and local sites and the beat reporters who actually break the news. They get scraped the same way, and they can’t afford to sue the AI companies.
Once they’re gone, AI will still be answering people’s questions. But where will it get the information to answer them?
Illuminating and thought-provoking piece. Thanks.