Guide

AI for local news: Five lessons from our first Lenfest AI Fellows gathering

By David Chivers

June 11, 2025

As AI adoption accelerates across the media industry, one of the most important questions remains: How do we ensure small and mid-sized newsrooms — not just the largest players — can benefit from these developments and leverage AI meaningfully?

In March, the Lenfest AI Collaborative and Fellowship Program – in collaboration with OpenAI and Microsoft – brought together our fellows, grantee organization executives, AI practitioners, researchers, and academics for our first in-person convening at The Walter Cronkite School of Journalism at Arizona State University. Here are five lessons we learned about building a more inclusive and effective AI ecosystem for local journalism:

1. Peer learning unlocks the most meaningful progress

Throughout the convening, a striking theme emerged: Peer-to-peer learning drove the deepest engagement.

Hearing journalists and technologists like those from The Minnesota Star Tribune and The Philadelphia Inquirer candidly discuss their AI product journeys demystified implementation and inspired attendees to accelerate their own AI adoption plans.  

• The Minnesota Star Tribune shared its Agate AI framework along with some of its early experiments and challenges as it innovates on its AI infrastructure. The team quickly recognized that reorganizing existing content around geographic markers could enhance local relevance and deliver greater value to readers, building on the organization’s statewide ambitions. 
• The Philadelphia Inquirer shared the latest iteration of its AI-powered “Research Assistant” to help journalists quickly access and synthesize decades of archival reporting, using semantic search and natural language queries to streamline research, improve citation accuracy, and unlock the value of historical content. The tool aims to strengthen reporting, save time, and serve as a model for how AI can support local reporting while preserving editorial standards. You can learn more about how The Inquirer built the tool here. 

Lesson: Technical skills matter — but trust, shared experiences, and community matter even more. Structured peer cohorts and case-study exchanges are more effective than top-down instruction alone.

How The Philadelphia Inquirer built an AI research assistant How The Philadelphia Inquirer built an AI research assistant

2. You don’t need engineers to innovate

Several successful local news AI experiments have come from individuals without technical backgrounds: Speaking at the convening, Dorinne Mendoza, the product and partnership lead of The American Journalism Project’s Product & AI Studio, highlighted how important it can be to have newsroom partners willing to experiment. A couple of examples she shared were: 

• Centro de Periodismo Investigativo in Puerto Rico used OpenAI’s API to build a custom translation tool — without a single full-time developer.
• Boyle Heights Beat, a neighborhood publication in Los Angeles, quickly adopted AI for wildfire communication, starting from scratch with simple prompts and templates.

Lesson: Curiosity, not coding, is the first requirement for newsroom AI innovation. With AI tools becoming increasingly more accessible, newsrooms — regardless of technical prowess or size — can quickly create tangible value with the simplest of tools. 

3. “Small wins” can have outsized impact

Group discussions highlighted projects where seemingly modest applications of AI created major ripple effects and/or build outsized momentum within your organization:

• Chicago Public Media’s AI fellow created a Slack-integrated AI tool to summarize traffic updates for WBEZ on-air hosts, saving time for the host and representing a lightweight utility that had immediate impact with minimal complexity. 
• CalMatters translated emergency information during wildfires, which allowed the outlet to reach more Californians who needed immediate help.
• San Antonio Report streamlined a decade-old sales process using AI-generated slides.

Lesson: Local publishers should prioritize tools that solve real, everyday problems — both in the newsroom and on their business teams. These “small wins” can catalyze broader newsroom transformation and unlock revenue growth.

4. Vendor evaluation tools are critically needed

Several news leaders said they’re inundated with pitches from AI vendors, tools, and startups — many of which promise transformative results. Small organizations, however, often lack the internal capacity to vet these vendors effectively, leading to wasted time and resources.

Lesson: There is an urgent need for clear vendor evaluation templates, landscape analyses, and trusted peer-reviewed directories to help news organizations make informed decisions about AI adoption. To get started, the Center for Cooperative Media created a free resource for publishers that recommends AI tools based on what tasks they are best suited for. 

At The Lenfest Institute, we’re working to support our cohort with practical tools — and, through partnerships with Arizona State University and others, extend these solutions to benefit the broader journalism industry.

5. Seek to learn from innovators in other industries

Engaging with AI innovators outside of journalism drives meaningful idea generation and opportunities to learn from novel approaches in other industries. For instance, many participants in our March gathering learned much from Arizona State University’s work in “principled innovation.” Ted Cross, an executive director at ASU, collaborates with university leaders to embed principles of integrity, curiosity, and civic purpose to help shape a university culture focused on human flourishing. 

The institution’s framework of “principled innovation” centers on asking not only what works, but what’s right, for whom, and at what cost. From designing more inclusive interfaces to asking how tech reshapes social norms (like replacing human interaction with automation), ASU urges its teams to pause, reflect, and map the full ecosystem of stakeholders impacted. 

For news organizations, the parallel is clear. If we build AI tools that only optimize for efficiency or cost reduction, we may unintentionally undermine equity, trust, or community relevance. As ASU emphasized, the real innovation lies beyond what AI simply can do. Instead, whether it advances human dignity, community connection, and democratic accountability. Sound familiar? 

Lesson: Innovation must be grounded in values, not just capabilities. Once rooted in shared principles, we can adapt and transfer breakthroughs from other industries in service to journalism’s mission. 

The path forward: From early cohorts to an industry ecosystem

The breakthroughs happening now — from civic transcription to multilingual reporting — point to a future where AI can strengthen and expand local journalism’s reach and resilience, not replace it. For local newsrooms looking to integrate AI, our Fellows have demonstrated that the path forward involves:  

• Starting small.
• Solving real problems.
• Centering the mission.
• Sharing what works (and what doesn’t).

Perhaps the most important insight from this convening is that early experiments are laying the groundwork for a more robust and resilient AI-journalism ecosystem.

But this requires sustained effort. 

At The Lenfest Institute AI Collaborative and Fellowship Program, our aim is to support projects that create scalable, ethical models for AI adoption — tools that can strengthen financial sustainability in local news and be rapidly adopted by peers. We’re also deepening co-development of AI solutions that engage new audiences, unlock revenue opportunities, and enhance news organizations’ efficiency through personalization and augmentation, all while upholding the core values of public-service journalism. 

This AI convening was just a first step. As we continue fostering this burgeoning AI-journalism ecosystem alongside our partners, we hope to advance the following priorities:

• Publish accessible case studies and playbooks showcasing practical applications of AI — including open-source code when possible — to make the projects as easy as possible to replicate.
• Create bridges between AI-focused innovation programs and broader journalism networks.
• Design lightweight, adaptable AI tools that small and mid-sized organizations can easily integrate into their workflows.
• Invest in peer-to-peer convenings and regional cohorts to extend knowledge beyond early adopters.

To learn more about the work emerging from The Lenfest Institute AI Collaborative and Fellowship Program — or to share your own perspective on AI in local news — contact David Chivers at  [email protected].