Our renewed $5M+ partnership with OpenAI will support further innovation, additional experimentation, and advance a new lane in product development.
Local newsrooms exist to serve their communities by keeping residents informed and engaged; today, they’re being asked to do this in a rapidly changing technological and information landscape. The American Journalism Project helps nonprofit local newsrooms grow and thrive, and our Product & AI Studio has become an increasingly important part of that work, demonstrating that emerging technologies and generative AI can be applied responsibly, practically, and in service of stronger journalism.
We launched the Studio in 2023, shortly after ChatGPT’s viral debut, with $5 million in anchor funding from OpenAI. The Patrick J. McGovern Foundation has also been a key partner in advancing this work, contributing a cumulative $1.6 million.
At the time, the questions around AI’s role in news were still wide open: where could AI actually help local newsrooms, where might it fall short, and what would responsible adoption look like in practice? Today, the answers are clearer. Across our portfolio, newsrooms are using AI to save time, expand civic information, strengthen fundraising, deepen reporting, and serve communities in more languages and formats, always with humans firmly in the lead.
Now, OpenAI is renewing its investment, contributing an additional $5 million over the next two years to help the Studio move from experimentation to field-building— most notably, by adding a new pillar focused on product development. In addition, OpenAI is providing API and Enterprise credits to support product innovation, and expanded access to its suite of Enterprise products to the entire AJP portfolio.
After two years focused on experimentation and identifying key use cases, we’ve seen tangible value for local news organizations as measured by dollars raised, time saved, and audiences reached. In our third year, we’re evolving the Studio to build reusable tools and infrastructure. These tools will first serve our portfolio of news organizations, which together reach more than 100 communities across 38 states, and ultimately, we hope to make them available across the broader nonprofit local news sector.
Key lessons have emerged from the past two years that are informing our Studio strategy moving forward:
What we’ve learned
AI is as much a leadership opportunity as it is a technical one. While human-in-the-loop AI systems have long been an intentional model for system application, we’ve begun using the phrase “human in the lead” to center the importance of strong leadership. The newsrooms making the most meaningful progress are those with engaged executives and teams, room to experiment, clear AI usage policies, and a commitment to embedding AI into actual workflows rather than parallel projects.
Start with real problems, not abstract possibilities. The most effective use cases came from identifying specific areas where teams were already stretched — translation, donor communications, meeting coverage, data analysis — and then identifying specific intervention points where LLMs could help.
Efficiency is a means, not an end. For lean news organizations, time savings are genuinely impactful. When teams save time on translation, campaign drafting, or meeting coverage, they unlock capacity for impactful work.
There is no single path to adoption. Some newsrooms build custom tools, others adapt off-the-shelf products, some integrate LLMs into existing back-end editorial and business workflows, and others use it to create new ones. Organizations are finding the use cases and workflows that work best for them.
From experimentation to impact
These lessons were forged over two years of hands-on work across the portfolio. To date, we’ve invested millions in AI projects across our portfolio, partnered with dozens of newsrooms on hands-on experimentation and shared our results on our website, through webinars and at conferences.
When we began our work, the Studio’s focus was experimentation and fluency — helping newsroom leaders understand emerging AI tools through hands-on testing and identifying where the technology could meaningfully support their work. Organizations experimented across a range of applications, from fundraising and translation workflows to civic information tools.
Some of the most meaningful results came from projects that could be adapted and reused across organizations.
Centro de Periodismo Investigativo’s translation work is a powerful example of how learning travels. They developed a clear workflow through LLM prompts and human review, which ensured both rapid and quality translations. The prompts contained details such as the audience, tone, and style guidance compliance, along with instructions to clarify cultural references and idioms.
Their approach was later adopted by Enlace Latino NC to produce its first English-language newsletter. “I was surprised how simple the whole process turned out to be,” said Nicolás Baintrub, newsletter editor for Enlace Latino NC. “I’m not a programmer, I’m just a journalist, a writer. I still was able to build this workflow just by talking with ChatGPT in plain language, almost like having a conversation with another person. And the results were better than I expected.”
That newsletter exceeded subscription goals and opened new sponsorship opportunities. Later, Boyle Heights Beat deployed the same translation model to deliver critical, real-time information during the L.A. fires— coverage that the newsroom could not have produced so quickly on its own. In a fast-moving emergency, this AI integration meant that multilingual communities had access to the trusted, reliable reporting they needed.
Later, the focus shifted to supporting cohorts in two areas where AI showed strong potential: revenue growth and unlocking civic information. In one standout revenue growth example, four newsrooms used AI to optimize their giving campaign workflow with support from BlueLena, and saw a 30% increase in end-of-year donations. On the civic information side, Wisconsin Watch dramatically increased its civic information output by using Gigafact Parser to identify claims made on podcasts, radio shows, and social media for its “Fact Briefs”— short pieces that counter trending misinformation, with an estimated 6x time savings per publication. Similarly, Civic News Company leveraged AI to transcribe, summarize and analyze school-related public meetings in New York, while City Bureau is using LLMs to surface patterns and synthesize observations across its rich repository of civic meetings.
Across both focus areas, AI confidence among cohort participants grew by roughly 30%, including the ability to identify use cases, teach others and deploy the technology responsibly. Newsrooms also built the leadership proficiency needed to develop effective AI usage policies and make strategic and informed decisions about when and how to use AI, due in part to our AI Leadership Series, ensuring that executives, not just practitioners, are equipped to lead their organizations through this moment.
What’s next
The Studio is evolving around three pillars: experimenting, sharing, and building. The first two have defined our work from the start. Now we’re adding a third — product development — taking what we’ve learned and turning it into tools that can scale for our portfolio.
- Building: We’re entering a new phase of product development, broadening our remit from isolated use cases to building capacity at scale, engineering reusable tools and infrastructure that have the potential to benefit the portfolio and, over time, the entire local news field. To start, we’re developing a tool purpose-built for revenue growth, designed to help nonprofit local news organizations enhance their communications with current and prospective donors. We’ll continue to identify and scale other high-impact tools to help local newsrooms thrive.
- Experimenting: We’re continuing to support news organizations in our portfolio to prototype their own tools and workflows that have real potential to help them achieve their goals and to scale across other newsrooms.
- Sharing: We’re expanding our knowledge-sharing practice through playbooks, case studies, peer learning opportunities, and a new resource hub for our portfolio. Our recently released playbook on using AI to sharpen newsroom messaging is one example, and many more are in progress, including deeper analyses of the examples shared above.
This next phase is made possible by OpenAI’s continued support: a second $5 million investment, alongside $3 million in OpenAI service credits, that will help us deepen direct support for news organizations while building reusable infrastructure and AI products for the field. The funding will support direct investments in additional newsroom partners, product development and experimentation, and expanded access to AI tools across our portfolio, including providing the nonprofit news organizations in our portfolio with optional access to ChatGPT Enterprise and API.
Over the past two years, we’ve watched newsrooms move from curiosity to confidence to capability and become more efficient as a result. That efficiency has unlocked new creative thinking around audience and revenue growth, along with new opportunities for engaging with civic information. We’re still early. But this work is a strategic imperative, equipping local news organizations to be strong, sustainable, and adaptable in our evolving information landscape. Our Product & AI Studio is committed to helping them build the technological capacity needed to serve their communities now and into the future.
For more information about the Product & AI Studio or to explore partnership opportunities, contact us at ajpstudio@theajp.org.
