Can the New York Times Rescue Journalism from AI Dominance?

Absolutely. I completely agree with you on that point.

While I’ve never engaged in it myself, I can imagine that when I reflect on all the investigative journalism we’ve conducted—often requiring months or even years to produce just one story—it becomes quite challenging to justify the expenses involved, doesn’t it?

However, I believe the public is often unaware of how much organizations like ours consider the public good. I truly share the conviction of our Founding Fathers, regardless of their differences, that a free society requires a free press, and that a nation aiming to navigate its own course must have access to independent and trustworthy information.

Jefferson famously declared that if faced with the choice of a government without press or a press without government, he would choose the press without a second thought. This is because this work is fundamentally important to society.

Journalism is, quite simply, the foundation of what we know. Transitioning to AI, it serves as the basis for how AI understands information as well.

Yes, let’s delve into that. Regarding AI in journalism, I categorize it into two areas. For WIRED, there’s the reporting on this pivotal moment in technology, which is crucial. Then, there’s the distinct inquiry, which we also address, on how AI will influence the practice and consumption of journalism.

The challenges here are significant, and I’m interested in how you view this at The Times.

My philosophy leans towards the idea that being first is rarely essential. There’s a danger in rushing into an emerging technology, and often, the risks of waiting—observing, learning, studying, and experimenting quietly—are less pronounced. Take a look at the news organizations that, in my opinion, quite foolishly chose to relinquish full control of writing and publishing articles to AI. Errors occurred.

Errors indeed occurred. And those errors were published.

That’s correct. It wasn’t just that mistakes were made; entirely fabricated journalists were created. This was an entirely foreseeable—and avoidable—series of blunders. Humans are accountable for everything we produce, end of story. There’s something about AI that has led some to forget that it is merely a tool.

Well, it does make for a very engaging conversation partner.

When my colleagues pose the question, “What if AI makes a mistake?” I respond, “Then it’s your responsibility to identify and remove that mistake from your work.” This wouldn’t even be questioned if we were talking about a computer. What if a computer makes a mistake?

You still have to catch it, right? I see this as a tool for reporting. Therefore, we utilize it more in the realms of reporting rather than writing or editing.

AI excels at recognizing patterns and anomalies within vast data sets. In many ways, this encapsulates a significant portion of reporting. For instance, campaign finance reporting essentially involves identifying patterns and understanding shifts in the mix of donors to both Republicans and Democrats, including large and small donors, along with any anomalies.