Signal & Trust – Issue VIII

signal scan

A curated round-up of how AI is reshaping news – and trust

New York Times Tests Out AI-generated Search Summaries

The Times is reportedly moving to test AI-generated story summaries, a move that has opened a rift with its editorial union.

  Source: Semafor

USA Today Co. Unions Call on the Company to Cut Ties with Palantir

Journalists from across 31 USA Today Company newsrooms say the deal poses an inherent conflict of interest – and threatens reader online safety and trust in the company.

Source: Poynter

From Guidelines to Architecture: How Newsrooms Are Rethinking AI Governance

As “keep a human in the loop” strains under volume of change, media managers from thirteen countries say they’re being forced to move AI governance from simple guidelines into systems.

Source: Reuters Institute

The Arena Group is Renaming Itself Paradium.ai and Its Revenue Just Halved

The publisher of The Street and Men’s Journal is rebranding and pivoting to an “asset-light” AI model as it responds to a steep revenue collapse.

Source: The Next Web

Cloudflare says Bot Blocking is Fuelling Publisher AI Deals

The company’s chief strategy officer says blocking AI crawlers is creating scarcity and leverage for publishers – including local and niche content providers – to negotiate from a position of strength.

Source: Press Gazette

“AI gives us a way to turn our journalism and editorial judgment into structured intelligence at a scale that was previously impractical. We can serve our audience that data, build new proprietary datasets, and use what we know about our own users to customize their experience with Politico.”

-Francesca Barber, Executive Vice President of Product, Audience and AI, Politico

Politico is betting that as commodity information proliferates, original facts and proprietary data only grow more valuable. Francesca Barber, Politico’s Executive Vice President of Product, Audience and AI, speaks with Aspen Digital about that bet, and what agentic AI could mean for how news gets produced and consumed.  This discussion has been edited for clarity and length.

Aspen Digital: Your career has moved across journalism, product, and technology — the New York Times, Google, a travel startup, and now leading AI and product strategy at Politico. How does that diverse background shape the way you think about what a news organization needs today?

Francesca Barber:  The common thread across my career has been sitting at the intersection of product and media. That is essentially the job I do now for Politico, overseeing product, design, audience, and AI. It comes down to knowing how to leverage technology and keep pace with innovation while staying grounded in the mission of journalism. For me, that mission is the responsibility of delivering original reporting and facts that people and institutions can act on.

My background in Silicon Valley is incredibly useful right now: not only do I have relationships with many of the innovators and disruptors who are sharing their insights and experiences with me, but I also feel I’m getting to go back to my old instincts of building and experimenting.

Aspen Digital: When Politico announced your expanded role last year, it made clear the stakes for the company were high: “AI is reshaping how professionals seek, filter, and act on information.” That’s a different value proposition than most news organizations. How does serving a professional policy audience change how AI innovation gets applied and what it has to deliver?

Francesca Barber: Politico is different from most news organizations in that not only do we have readers who are hobbyists, but a lot of our paying Politico Pro subscribers rely on us for their daily decision-making at work.  That undoubtedly influences our AI vision. It means we must provide insightful, accurate, and essential information and data.

For me, the most exciting part of this moment is that AI is all about data, and we are a data intelligence and journalism business. Internally, we have the opportunity to make faster decisions and build more efficient workflows, and that is important. But the bigger opportunity is for our readers, subscribers, and clients, who need our journalism and data throughout the day to make decisions for their jobs.  In this context, data means the public record like hearings and filings, and, critically, the proprietary intelligence our reporters generate as a byproduct of their reporting. Take Influence, our daily read on who and what is moving where across the policy world: we can turn that journalism into a living database of people and organizations, linked to public datasets, so AI surfaces signals and patterns competitors cannot see. AI gives us a way to turn our journalism and editorial judgment into structured intelligence at a scale that was previously impractical. We can serve our audience that data, build new proprietary datasets, and use what we know about our own users to customize their experience with Politico.

Aspen Digital: You were given an explicit mandate from Politico leadership: to empower “every reporter, editor, and publishing professional to be more effective and creative.” One year later, what have you been able to develop and deploy to deliver on that – and what challenges have you encountered?

Francesca Barber: Rather than just asking questions of a chatbot, our reporters and editors are spending time thinking deeply about what problems AI can solve for them and then, ultimately, building the solutions themselves.

We launched a Newsroom Innovation Hub this summer to make that systematic. The Hub builds AI-powered workflows and tools that change how our newsroom reports, edits, and publishes. Since June, the Hub has also been running hackathons that turn new AI capabilities into practical workflows. In this process, we have built more than 15 prototypes and more than 10 working tools. Some of my favorite moments have been watching people who had never used AI create a working tool in a single session, then leave excited to keep experimenting or even share what they built with colleagues.

Our Brussels Playbook is a good example. In one session the team built four tools: one assembles the next day’s calendar and news into a Slack briefing, another checks names and job titles in copy with supporting sources, and two monitor EU sources in the background and alert the team when something relevant lands, like new European Commission documents. Less time is spent on repetitive gathering and checking, and that means more time can be spent on original and distinctive reporting.

Aspen Digital: Politico Pro has tens of thousands of subscribers who rely on it for policy intelligence. But AI tools like ChatGPT and Claude can, in theory, now answer many of the same questions your subscribers might previously have relied on Politico to deliver. Are you seeing audience behavior shift in that direction?

Francesca Barber: Many of the data sources on Politico Pro are not publicly accessible, so a general AI tool cannot reproduce them. More fundamentally, raw public data and AI alone can never match what journalists gather. Our value comes from Politico defining the journalistic question, identifying the right material, and structuring it in a useful way. AI makes that feasible at scale, and we’re taking advantage of that. But the intelligence itself remains a product of human judgment, which is why a data-only competitor cannot easily replicate it.

For example, in our mandatory detention investigation, AI helped us extract and standardize information from more than 13,000 court rulings, turning fragmented public records into a proprietary dataset that powered both a major investigation and a searchable resource. We leveraged AI, paired it with our journalistic judgment, and produced reporting nobody else had. That’s not accessible through AI platforms.

At the same time, we are thinking about how to meet subscribers where they are. And yes, in some places, they are moving from destination habits to workflow integration, and they want trusted, citable intelligence inside the tools they already use. This has been particularly true of our European subscribers. So rather than fight that shift, we are working to deliver Politico journalism and data directly into those workflows.

Aspen Digital: Agentic AI – systems that act autonomously on a user’s behalf, rather than simply answering questions – may be the next major disruption for news organizations. How do you expect it to change the way news is consumed, and the way it’s produced?

Francesca Barber: This is one of the most exciting moments in my career – and I really believe, if we maintain our mission and sense of focus, we can figure out how these new technological changes open up a whole new world for the media. Our industry creates original facts and data sets. That is a scarcity in a world that is filled with commodity information. So there is a real opportunity for us to think about how to leverage that, reimagine our formats, and experiment with new ways to deliver essential information.

Think about how consumer habits have evolved. For a long time we only read the news. Then we listened to it, then we watched it. Now, on top of all of that, people also ask, create, and build. Our job is to meet users wherever they are, give them flexible ways to engage, and opportunities to continue to evolve their habits. There is real value in an editorially curated media experience, and there is real value in customization, because everything is about to get hyper-niche now that anyone can build something for themselves.