By Katie Drummond

Business

Jul 28, 2026 6:30 AM

Can the New York Times Save Journalism From Our AI Overlords?

In 2023, the Times sued OpenAI and Microsoft for copyright infringement. They’ve since spent more than $20 million on the case, and publisher A.G. Sulzberger has no plans to stop fighting it.

---

---

For most companies, spending upwards of $20 million on a single legal fight with an uncertain outcome might mean the boss has some explaining to do. For A.G. Sulzberger, the publisher of The New York Times, that explanation is simple: It’s the cost of defending journalism—at the Times and everywhere else.

Sulzberger, who took on the publisher role about a decade ago and has been an influential leader in the Times’ much-lauded push into digital innovation, has no shortage of defending to do lately. Nearly three years ago, the Times sued OpenAI and Microsoft for copyright infringement, alleging that the companies illegally used New York Times journalism to train AI models. That legal case, in which motions are still being filed about discovery requests, is only one front in a broader battle: The Trump administration is waging an ongoing war on press freedoms—most recently with an outrageous attempt to subpoena Times journalists over coverage of the president’s Qatar-gifted jet to be used as Air Force One—while AI companies radically reshape the way people find and consume information online.

The Times has arguably never been stronger as a business, having recently surpassed 13 million subscribers, and its newsroom is widely seen among journalists as one of the most stable and well-resourced in the business. Still, Sulzberger believes that journalism itself is facing myriad existential dilemmas—from an administration ready and willing to target reporters, from AI companies whose products depend on the very reporting they risk undermining, and media companies being demolished by the changes they’re unleashing on the open web.

I sat down with Sulzberger in late July at WIRED’s offices. We talked about that Times lawsuit, how and where AI is transforming reporting inside the newsroom, and why he thinks journalists and media executives need to team up—even while they continue to compete fiercely. You can read our entire conversation below, or listen to it here.

This interview has been edited for length and clarity.

KATIE DRUMMOND: AG, welcome. Thank you for being here.

A.G. SULZBERGER: It’s great to be here. Thanks for inviting me.

I want to get into AI, and I will, but I would be a terrible journalist if I didn’t start with news that has recently affected your paper.

Specifically, I’m talking about the subpoenas that the Justice Department has issued to several Times journalists about their reporting on the Air Force One jet that was gifted to the Trump administration by Qatar. The Justice Department also issued subpoenas to third-party phone service providers to obtain text messages and phone calls between some of those journalists and their family members.

Thanks for starting there, because it’s obviously yet another disturbing escalation in the Trump administration’s efforts to punish and undermine the free press in this country.

First I should say, I have so much admiration for the reporters who have done this work and who are doing it under a different kind of pressure than we have seen in generations. I think that by any objective metric, this is the biggest effort to go after the press that we’ve seen from a White House probably since Woodrow Wilson, so in more than a century.

These reporters just continued, day after day, to put their heads down, right? Not to get baited by the folks booing them or the folks cheering them on, and just to do their work. That work includes reporting things that the President of the United States doesn’t want the public to know because it may be unflattering or embarrassing.

In this case, they reported a story, as you said, on the Qatari-donated jet, and it’s a little unclear: Was it donated to the Trump administration or to Trump himself? A donation worth hundreds of millions of dollars.

Right.

When that occurred, a lot of people had questions about whether it could ever be secure enough to hold the President of the United States and ferry him around the world. Sure enough, our reporters found out that on one of its inaugural trips, its first trip overseas to Turkey, there were concerns that it was insufficiently equipped to respond to a threat, that the president had to fly back on the original Air Force One. He denied this initially. But we exposed the truth, which has subsequently been confirmed.

Then it gets even stranger.

Every administration, we will publish something that the President of the United States doesn’t want the American people to know. Upsetting a president of either party is nothing new to us. What is new is that the administration defied many of its own standards and practices and explicit guidelines to very rapidly, in the course of a day or two, serve subpoenas on these reporters. They went straight to the homes of reporters, in an obvious attempt to intimidate them. As you said, we are now learning that they sought to get records, not just from the reporters themselves, but from members of their families. It’s a gross overreach.

Just to contextualize for listeners, how unusual is all of that? To have law enforcement at your home as a journalist, to be subpoenaed right out of the gate, to have those phone records called up by the Justice Department. How weird is that?

I mean, it’s extremely unusual. It’s very rare. The closest comparison that I have had at the Times in my tenure—I’ve been in this job roughly a decade—happened in the first Trump administration, when some of our reporter records were secretly seized by the Justice Department. In fact, that led to some reforms in the Justice Department aimed at curtailing that type of abuse.

On one side this is objectively very rare—there was similar overreach, you may remember, in a Washington Post case some months ago, where they actually raided the reporter’s home and seized her devices—[but] on the other [hand], this is building naturally off of what I think is a broader effort.

You have also criticized other media outlets for capitulating to the administration, particularly around settling what you’ve called winnable cases to appease Trump. Do you still see your peers as largely capitulating?

I don’t think I ever said largely capitulating. I actually think that the industry has largely held its ground. And to be very clear, holding your ground does not mean an adversarial posture. It means being willing to publish the truth even when powerful people get angry at you for publishing it.

It means being willing to ask any question, no matter who it might upset. I actually think that the industry, by and large, has done very good work. WIRED’s done good work. The [Wall Street] Journal has done terrific work. The Associated Press. When the White House conditioned their continued access to the White House on them switching from calling a big body of water the Gulf of America instead of the Gulf of Mexico, they didn’t take that sitting down.

I remember this.

They sued, and I think rightly. They sued regardless of whether they win the case, right?

Right.

One of the things that I’m saying to peers in the industry regularly is that your rights only hold if you seek to enforce them.

You may not win every battle, but the surest way to lose your rights is to avert your eyes, drop your head, and try to avoid notice and just take what comes.

I do think that some news organizations have clearly felt the pressure from the administration and I think allowed that pressure to affect coverage.

My colleagues in Washington tell me that they’re regularly talking to reporters at other news organizations who are saying that they will no longer pursue certain stories, because they’re too worried that their editors will kill them because of fear of retaliation, and that their model is actually to wait for an institution like the Times to report the story, and then they’ll follow it, and they think that that’s a fight they can win. But breaking the story, being the one to first upset the White House, seems like a road too far.

That is the definition of chilling, right? That’s when the message that retaliation and punishment must be avoided gets down to the lowest levels of the newsroom. Now, my job as publisher of the Times is to make sure the opposite message is getting through, right?

After these great reporters were served these subpoenas, I called each of them and I said, “We have your back, and we’ll do whatever we need to support you.” One of the things that I was most encouraged to hear in those conversations, although not surprised, these are tough guys, was that they said, “Oh, we’re not gonna let go of this story.”

Good. As a journalist myself, and I run a publication that takes heat here and there, I worry sometimes about how we better translate that to the audience. So making sure the public understands how essential this work is, and exposing them to the work itself, especially in the context of AI. How do you think about that? What else can news organizations be doing to make sure that audiences are exposed to the work and see and understand the value of the work, and what it means when that work stops happening?

Look, I agree that we haven’t done good enough of a job as an industry on that. We’ve sort of taken it for granted. At the Times I think about just the outsize investment we are making on behalf of the public’s right to know. We put reporters on the ground in 150 countries last year. Every year, over 150 countries. Last year, we had reporters on the ground in Myanmar, in Sudan, in Lebanon, in Iran, in Venezuela, in Gaza, in Israel, in Russia.

In Ukraine alone, we had more than 70 journalists on the ground. And I am sure, although we’ve never done it, that if we dug in, it’s really hard to find the financial case for those expenditures.

Yes. I would agree with you there.

I am sure, though I’ve never done it, that when I look at all the investigative reporting we’ve done, and this is reporting that takes months, sometimes years of work to produce a single story, that it would be very hard to justify all those expenditures, right?

But I think the public sometimes doesn’t know how much institutions like ours think about the civic good. I really believe, as our Founding Fathers did, and across all their lines of division, the Founding Fathers agreed that a free people need a free press and that a country that aimed to steer itself needed access to independent and reliable information.

Jefferson famously said, given the choice between a government without a press or a press without the government, he wouldn’t hesitate to choose the press. And that’s because this work really does matter to society.

Journalism is quite simply very often how you know what you know. Shifting over to AI, it’s how AI knows what it knows.

Yes, let’s get into that. In the context of AI in journalism, I break it into two buckets. For WIRED, there’s reporting on this technological moment in time, which is really important. And then there’s the separate question, although we cover that as well, of how AI will affect the practice of journalism and the consumption of journalism.

I mean, there are really difficult lines to walk here, and I’m curious about how you think about that at The Times.

I think my philosophy is you very rarely in life need to be first. There is a risk of jumping too quickly into an emerging technology. And there’s probably more risk in that than waiting, watching, learning, studying, experimenting quietly behind the scenes. I think you saw some of those risks, right? Like, think about the news organizations that, in my mind incredibly foolishly, decided to just hand over full responsibility for writing and publishing articles to AI. Mistakes were made.

Mistakes were made. And mistakes were published.

That is right. And not just mistakes were published, but fake journalists were invented. It was a completely foreseeable and unnecessary set of mistakes. Humans are responsible for everything we produce, period. And there’s something about AI that has tricked people into forgetting that it is a tool.

Well, it’s a very good conversation partner.

But when my colleagues say, “Well, what if AI makes a mistake?” I say, “Well, then it’s your job to find that mistake and take it out of your work.” You would never ask that about a computer. What if a computer makes a mistake?

Well, you gotta find it, right? I view this as a reporting tool. So we’re using it much more in reporting than writing, crafting, editing.

I think AI is very good at spotting patterns and anomalies in large data sets. In many ways, that’s a pretty good summation of a big part of reporting. You know, campaign finance reporting is basically what are the patterns, how is the mix of donors to the Republicans and Democrats changing, big donors for small donors, and then anomalies.

One of the reporters who won our first Pulitzer that involved AI-supported reporting, what he said to me is, “You have to think of AI as a filter.”

In his case, he was examining bomb craters in the Israel-Gaza conflict, and he said that it regularly makes a certain kind of mistake. So the way to think about AI is it took millions of images and it shrunk it down to thousands of images. But it was his job as a journalist to look at every one of those thousands of images to make sure he felt confident in the findings.

So that’s the type of experiment we’re doing, and I think between that natural posture of, we wanna take our time and move with curiosity and creativity, but also some degree of care and caution, and that we as journalists, as reporters, as editors, are ultimately responsible for the work.

I think that’s how we’ve avoided most mistakes.

I agree with you. I think it is a very valuable tool, or can be used to create very valuable tools, for reporting and for reporters, especially as newsrooms have gotten smaller. I think there’s a competitive advantage there. I will say at WIRED we are building our own AI tools for reporting, and one of the goals is actually to beat The New York Times more often. So we’re trying to get clever in ways that maybe other news organizations aren’t when it comes to using AI.

Everyone needs competition. So we’ll watch with interest.

But in a speech recently, you also said there are early signs to be concerned about when it comes to how AI will affect fact-based reporting. I mean, some of those are obvious. But talk to me about what you see as those early signs. What worries you most here?

It’s really important just to go back to what I referred to in that speech as the original sin of AI.

There are four ingredients in good AI products: There are the people who write the code; there is what we call compute, the chips and data centers, et cetera; there is electricity; and there is what Silicon Valley calls data but which is content, journalism, books, music, movies.

And let’s just go through those four right now. On the coders, you literally have contracts that are sports-star level.

Oh, they’re paid like NBA players. It’s outrageous.

Hundreds of millions of dollars for a single individual coder at times. Now let’s look at compute and electricity. Literally, these companies are taking hundreds of billions of dollars of debt in order to pay for those things. That’s how big the prize is if they win. These are companies that all now have trillion-dollar valuations or near-trillion-dollar valuations, right? So the prize is big.

Now you get to data, and they’ve just taken it. They’ve just taken it all. Without permission and without any compensation, because they thought they could. They think they can get away with it. And if they can get away with that, then that boosts the profit margins of the richest and most valuable companies in the history of humanity. And if the casualty is that it completely guts what remains of our already financially beleaguered industry of journalism, so be it.

Read more: Can the New York Times Save Journalism From Our AI Overlords? – WIRED

Continue/Read Original Source: Can the New York Times Save Journalism From Our AI Overlords? | WIRED