Newsletter

Context Window: Tools and news you can use for AI in local media

A new newsletter from the Lenfest AI Collaborative and Fellowship Program

September 28, 2026

Hello, everyone,

I’m Sonali Verma, and I am excited to share AI learnings and insights with you through the Lenfest Institute’s new newsletter, Context Window.

Why do we call it Context Window? For those unfamiliar, this is a term describing the amount of content a large language model can consider for any given query. It is defined fully here.

But for me, the name “context window” works on another level. As a journalist, machine-learning product manager and AI coach, I know context is what differentiates run-of-the-mill, widely available commodity information from higher-value output that actually meets user needs and solves real problems. It is vital for both journalism and AI.

This newsletter is not only going to be a window into some truly innovative AI work that is underway at local news organizations across the country, it will also serve as a resource hub: it will prioritize sharing ideas and tools that you can pick up and apply right away.

The Lenfest Institute just announced the next phase of the Lenfest AI Collaborative and Fellowship Program, thanks to support from OpenAI. Over the past couple of years, the first cohort of Lenfest AI Fellows have contemplated, debated, launched, adapted, and scaled AI projects. These engineers come from diverse backgrounds, with some working in the news business for the first time ever and some blazing their own trail through the newsroom.

They have a wealth of learnings to share with you – and we thought this newsletter would be a great way to do that.

Knowledge sharing is at the core of the Institute’s mission. We want news organizations to replicate good ideas and avoid bad ones. Everyone in the news business is on the same journey. Wouldn’t it be nice if we could spare each other the effort of reinventing the wheel?

Our inaugural edition kicks off with a quick look inside the Lenfest AI Collaborative. If a colleague forwarded this email to you, you can subscribe here.

Our first edition of Context Window includes three components:

An introduction to our Fellows and key projects you can use.

An interview with Tristan Loper of The Lenfest Institute, who oversees the fellowship.

And, as always, links to additional resources.

Our AI Fellows and their projects

Our AI Fellows are working on a plethora of ways to help news organizations build sustainable businesses. Here’s a very quick look at some of their projects:

Rajesh Barade, The Seattle Times

An AI-powered workflow that ingests, transcribes, and analyzes City Council video recordings to surface a customized set of newsroom-specific topics of interest.
Tools that support advertising prospecting, sales, and new revenue.

Aaron Brezel, ProPublica

Moving from simple chatbots to data-interfacing applications (e.g., chatting with a spreadsheet).

Dana Chiueh, ProPublica

Using local LLMs to process sensitive tips and a secure internal search engine for investigations.

Mark Chonofsky, Chicago Public Media

Mass transcription of decades of audio archives to make them accessible first to staff, then to the public.
Active English-to-Spanish translation pipelines for newsroom archives and daily reporting.

Kevin Delaney, The San Francisco Standard:

An AI-powered news app designed to reimagine how audiences interact with local information – a mobile interface that offers dynamic interfaces, letting readers interact with local reporting in real time.

EJ Fox, Arizona State University NEWSWELL

Automating ad and content layouts for print, specifically targeting efficiencies for hyperlocal papers, to significantly reduce overhead costs and increase resources for digital products.

Kevin Hoffman, The Philadelphia Inquirer

A local media aggregator that pulls news from hyperlocal sources for newsletter curation
Dewey, a digital librarian allowing natural language search of a half-century of archives

Christopher St. John, who takes over from Paulino Ramirez, The Dallas Morning News

An on-article feature that generates questions and links users directly to relevant paragraphs or related stories.

Peter Rasmussen, The Baltimore Banner

Developing an AI system to analyze the newsworthiness of state and local data.

Omar Reid, The Boston Globe

Brand safety tools to increase the value and targeting of digital advertising
Prospecting agents for the ad sales team

Ryan Thayer, The Minnesota Star Tribune

A chatbot that allows the newsroom to access archives.
Dynamic sentiment and entity extraction to power better content personalization and targeting.

Shanni You, Newsday (Long Island, New York)

Creating a centralized data hub, that exposes internal data for public queries and internal investigations.

Q&A with Tristan Loper, head of national programs at The Lenfest Institute:

Why are we even doing this? Why does it matter?

The Lenfest Institute has always cared about great journalism, but we also recognize that there is no journalism without a sustainable business behind it. That’s the lane we chose for the AI Fellowship, and I think it’s what makes the program distinctive.

From the start, we framed the program as a way to strengthen the business operations that support journalism – things like workflows, revenue, and audience growth – rather than a way to replace journalists’ expertise or cut jobs. It would be a travesty if, after all their training and experience, journalists were reduced to just editing AI-generated copy and checking it for hallucinations.

Instead, by focusing on business operations and embedding Fellows directly into newsrooms, we’ve shown there are many valuable use cases for AI that help news organizations grow without sacrificing quality or undermining the craft.

It’s been a couple of years since we launched this. What have you learned from it?

The biggest lesson for me is that when hands-on users drive technical solutions, the potential benefit is enormous – but when AI is imposed from the top down, the risks are just as great.

AI is more than just a large language model or a chatbot; it’s a broad toolkit that can solve a wide range of problems. But it only works well when you start with real people’s needs and constraints.

Our Fellows sit inside newsrooms, listening to colleagues talk about workflow headaches, business challenges, and even some journalistic opportunities, and then ask, “What are your problems, and how can I help you address them?”

This user-driven, people-centered approach respects people’s jobs and livelihoods and focuses on making their work easier, more effective, and closer to the wish-list projects they haven’t had time to tackle. It’s a very different posture from ‘you must use this tool now,’ and I think that difference has been crucial.

One other benefit we’ve seen stems from the program being platform-agnostic. Even though it is supported by OpenAI and Microsoft, our Fellows use the best technology they can find for their use cases. This flexibility has allowed them to approach problems from every angle.

What’s one thing you love about this initiative?

Over time, the Fellowship has evolved into a tight-knit, collaborative community, and that is just as important as any single project. The Fellows not only work inside their own newsrooms, they learn from and build on each other’s experiments. The Philadelphia Inquirer’s archive tool Dewey, for example, has been adapted at least three times by other members of the Fellowship.

My hope is that we can extend that culture of shared learning and user-first thinking well beyond our grantees, so it becomes a permanent trend in how newsrooms approach AI. If we get this right, AI won’t be something done to journalists from the outside; it will be something shaped by the people who understand their audiences, their communities, and their own work best. That’s how we protect journalism while building stronger, more resilient news businesses.

Looking ahead

Here is a sneak peek into what we’re planning in the months ahead.

Learnings from the 2nd anniversary convening: Who doesn’t love cake? We just celebrated two years of this immensely successful program by bringing the Fellows together to share learnings. Keep an eye on your inbox for key takeaways from the event.

Where else will you find our Fellows sharing their insights?