Value Creation in the Age of AI | Interview with Florent Daudens

In the latest episode of AI Frontrunners in News, our podcast series exploring how artificial intelligence is reshaping the architecture of journalism, we spoke with Florent Daudens, co-founder of Mizal AI and former Press Lead at Hugging Face.

Florent has spent years inside newsrooms, including at CBC/Radio-Canada, and Le Devoir. He also lectures at Université de Montreal and CUNY. These experiences in tech companies, academia, and publishing groups gives him a rare vantage point on how technology is transforming media from the inside out.

This interview has been lightly edited for clarity and readability. Listen to the full conversation on Spotify and YouTube.

Over the past couple of weeks, we’ve seen a lot of talk around Moltbook and OpenClaw. How do you look at this phenomenon?

I think it’s both weird and fun.

It can be a good thing because it brings to the general public a not-so-distant future where agents collaborating together becomes tangible, so at least people can start to think about it.

At the end of the day, we will be in a world where agents mediate a lot of information and take actions for us. So how do you evaluate, first of all, that it’s really an agent? How do you evaluate the level of trust you can have in this agent? Where does its expertise come from? What are its goals? These are some of the questions it helps us think through.

What is the relevance of OpenClaw/Moltbook for a newsroom leader?

The main takeaway for me is that agents will consume content much more than humans. We’re seeing data points that suggest they already consume more content than we do. The question becomes: where will value creation lie? Will it shift, or will it expand?

I think value creation will expand because if you are able to create content for agents’ consumption, you can create monetization at scale. On the other hand, it will also highlight the importance of brands and connections, because human connection will be more important than ever.

So how do you find the right balance between having a strong brand and bringing forward your individual voices (your internal journalists) and putting them in the spotlight so they can create more direct connections with their audiences?

On the technical side, things will be much more complicated because you will need to think about dual-format publishing: one for bots and one for human consumption. You really need to rethink your architecture. Today, I feel like we’re too focused on building experiments for humans, whereas we don’t think enough about how to architect content for machines — even though we can create value from that.

You’ve been in the hotbed of the open source ecosystem: Hugging Face. How can the news industry benefit from open source?

Working as the Press Lead at Hugging Face was like being at the epicenter of open source AI and trying to articulate how it can be valuable for the news industry. My main takeaway is that we need to collaborate more as an industry — not by just creating committees, but by sharing code and reading each other innovations.

You can compete on journalism, but not on the plumbing. When you build together, it accelerates innovation. You can focus on incremental productivity gains, but the real goal is to accelerate innovation and stay up to date with the pace of it within the broader AI ecosystem.

I agree that you don’t see much co-creation among publishers. Why do you think that is?

I would have two hypotheses. The first is that many news organizations see AI as highly competitive, so they might not want to collaborate on it. The second is that the overall technical skill level within the industry needs to be higher. If you ramp up internal skills, you may be more eager to share what you’ve learned from your experiments.

You’ve written about the risks facing mid-sized publishers. What advice do you have for them?

I would say there are two archetypes. You can go all in on premium, voice-led brands, where the key question becomes how you maintain a strong audience connection. People will be willing to pay for that, especially in an AI-media ecosystem, because they will want to sift through the noise.

The second possibility — which is more complex for a single mid-sized organization — is betting on becoming a distribution platform. You build the pipelines, the APIs, the fact-checking services, and so on.

In both cases, what’s interesting is thinking about how to unbundle the artifacts from the value you create. Where do we create value, and how can we distribute it? When you return to these two questions, you can have a much deeper conversation about where the future of each news organization actually lies.

Conclusion

If there is a common thread running through this discussion, it is this: AI will reshape where value is created, how content is consumed, and how news organizations position themselves in an increasingly agent-driven ecosystem.

Whether through premium voice-led brands, distribution infrastructure, or deeper industry collaboration, publishers will need to make deliberate choices about where they compete and where they collaborate.

To hear the full conversation, including additional examples and deeper reflections, listen to this episode of AI Frontrunners in News on Spotify or watch the full discussion on YouTube.