Exclusive Interview with Ezra Eeman: “Rethinking journalism in a fundamental way”

Ezra Eeman has spent the past two decades witnessing first-hand how technologies are reshaping newsrooms. Director of Strategy & Innovation at NPO, he has recently been appointed as the Director, AI in Media by WAN-IFRA, where he believes he can play an essential role in an international organization that connects “all the players in the markets”. 

I sat down with Ezra ahead of the Digital Growth Summit 2026 where he will challenge us to an “unplugged” discussion on AI to chat about how news discovery is reversing, why efficiency gains aren’t enough, and what publishers can do with the intelligence they already have.

The way people discover news now has changed dramatically. As Ezra explains, the “direction of the action” has reversed: instead of seeking out content, content is finding you through algorithms in your feed.  AI has sped up the enablement of this.  

What impact does will this have on publisher-owned products? “Convenience is very attractive in products,” Ezra shares, “if I open my smartphone and there are three articles that you should read now, that’s different than when I have to think about three topics I want to know about now. And so that’s a shift that you will see – it’s a removal of friction, it’s an anticipation of needs”.

“Current use cases [for AI] are based around how to go faster,” Ezra explains, but they are “not rethinking journalism in a fundamental way”. Indeed, the main uses for AI in news organizations today is to “speed the process”, but done so the constraints of the current business model: automating cumbersome workflows, speeding up data analysis, etc.

However, Ezra is adamant on the need to “rethink journalism itself”. Some open questions he gave as examples were: 

• What would it be like if you could use 10 agents instead of 10 journalists? Would the role of a journalist to be a sort of manager of the agents? What kind of reporting would this require?

• Is the article the best format to bring a story to an audience?

"We’re not yet transforming ourselves or pushing ourselves towards the next business model."

- Ezra Eeman, Strategy and Innovation Director

While visionary thinking on AI is certainly needed to go beyond what our current journalism business model is based on, Ezra also provides practical advice for publishers that could be used today.

First, use AI to know your audience better. “Your audience gives you a lot of signals about what they like and how they read. All of this is currently not yet brought together as a new kind of intelligence that the newsroom can use.”

On the flip side, knowing who your audience is key, but knowing how they use AI is crucial as well. “What are they trying to do? Can you help them with what they’re trying to do?”, are some of the starting questions Ezra recommends reflecting on. “Otherwise, you’re making a lot of assumptions out of thin-air.”

On synthetic audiences

Knowing your audience with AI has hit some limits. Ezra shares that experiments to build synthetic audiences using AI have achieved mixed results in practice. It was born out of the fact that newsrooms don’t represent complete reader demographics so the idea was to help journalists understand how different groups of people would have different perspectives on a story before publication. Yet the challenge was in representing each community accurately and the fact that “synthetic audiences do not evolve with real people. It doesn’t live.”

Second, develop a strong editorial intelligence. Leverage AI to help unpack the “so what” of a story. For instance, by using semantic search in your archives, you can uncover important relationships to the story and see how it connects to the ones published. AI can uncover, for example, the fact that last year 20 stories were published on wildfires whereas this year there have been 30, highlighting an interesting relationship.

Also consider the format of the story “I’d like to see how an article could become an explainer video, or how I can have a conversation with an AI-generated character that leads me to a story”. Animated about formats that combine conversational and interactive storytelling, Ezra highlights that these would work amazingly for explaining general concepts, while maintaining a natural tone.

As I reflect back on my conversation with Ezra, here’s are three questions I think publishers should be asking themselves:

• Do you know how your audience uses AI?

• What would 10 agents let you report that 10 journalists cannot?

• Is the article still the right container?

If you’re interested in discussing these questions and more, join us at the Digital Growth Summit on October 15th where Ezra and 130 other leaders in the news industry will explore how tech is shaping the future of news