Expanding our team and other recent developments: an update from our Product & AI Studio

When we launched the American Journalism Project’s Product & AI Studio in 2023, we were clear-eyed about our objective: to help the local news organizations in our portfolio understand and strategically leverage the rise of artificial intelligence and other emerging technologies. 

Since the launch of the Studio, which was made possible by OpenAI with additional support from the Patrick J. McGovern Foundation, we’ve continued to make progress. In December, we shared the first set of grantees who would receive support from the Studio. Today, we’re sharing more about the work we’re doing through the Studio, including the team we’ve assembled and some of the advancements the local news organizations in our portfolio are making.

Building our team

We’ve assembled a team of talented professionals with extensive  backgrounds in media technology and partnership development to lead the Studio. Together, they support our portfolio in experimenting with smart applications of technology to advance their newsrooms’ public service missions; they will also advance opportunities to share our learnings and findings across the local journalism field. 

  • Trei Brundrett is consulting on the development and direction of the Studio as senior adviser, bringing a wealth of experience in digital media. He began consulting with us in 2022 to advise our Startup Studio grantees on the adoption of technology and audience development. Last year, he led the development and launch of our Product & AI Studio, drawing on his expertise in news and revenue product development as the cofounder of Vox Media and the leader of their product, technology and data organization, as well as through his board service for media companies like The Texas Tribune and The Guardian.

  • Dorrine Mendoza is the product & partnerships lead for the Product & AI Studio; she began this role in March 2024. Dorrine is the American Journalism Project’s point person for understanding the needs of nonprofit news organizations, identifying new opportunities for impact with emerging technologies, and developing relationships with technology and platform partners, universities and other organizations exploring news product strategy. Dorrine was previously at the Local Media Association, where she coached local media organizations on sustainability. She was also part of Meta’s News Partnerships team, where she created and led its local news accelerator programs, and worked in social media at CNN. She began her career in local news.

  • Liam Andrew began as technology lead for the Product & AI Studio in May 2024. With an expertise in software and data engineering, Liam is responsible for understanding and helping our portfolio act upon existing and emerging technologies — specifically artificial intelligence — by researching, designing, developing and deploying solutions. Previously, he was chief product officer at The Texas Tribune. His background leading technology teams in nonprofit news and his knowledge of machine learning will help him act as a coach and collaborator for our portfolio as they experiment with AI and implement new products and technologies.

Experimenting with AI: Early insights

In December, we gave an update about the thirteen news organizations in our portfolio that were awarded grants to experiment with applications of AI in their organizations. All of these organizations are making meaningful progress by testing how AI — in particular, generative AI powered by large language models (LLMs) — is valuable in the journalism context, across a wide range of novel use cases that touch audience development, support existing reporting practices and increase membership revenue. 

This cohort is sharing lessons learned with each other, publishing their findings, and presenting their learnings at conferences. Our team is identifying common themes of opportunity and patterns of challenges to overcome, with the goal of sharing those with our full portfolio and the broader industry. 

We believe this type of intentional, hands-on experimentation and thoughtful sharing will help local nonprofit newsrooms better understand the potential and downsides of AI. This will also position them to take a leadership role in how the implementation of AI and other emerging technologies should evolve and best serve their communities. 

A few early highlights: 

  • THE CITY recently published an update on its use of ChatGPT to map stories across New York City. The approach utilized AI to enhance the reach and impact of its reporting, making it easier for residents to navigate and engage with local news. You can read more about the work on THE CITY’s website.
  • Spotlight PA developed an AI-powered election assistant in collaboration with Dewey Labs. Days before Pennsylvania’s 2024 primary election, 200 Spotlight PA readers were invited to test this tool, designed to answer common election-related questions. By leveraging AI to provide immediate, accurate responses, the tool aimed to enhance voter engagement and accessibility.
  • The Marshall Project recently hosted a workshop on large language model prompting techniques for reporting tasks. In it, the news organization explored common mistakes and concerns around using LLMs for news and the mechanics of how they work. The Marshall Project also provided a series of hands-on exercises to teach different ways to interact with chat interfaces to LLMs to inform reporting and released a living tip sheet.
  • When Sahan Journal wanted to better understand how Minnesota’s charter schools award contracts, the outlet turned to AI to help process more than 170 documents. More than two-thirds of students at Minnesota charter schools are people of color, which makes these schools an especially important topic for the outlet, but charter-school governance and spending often lack transparency. Sahan requested nonprofit tax filings from almost every charter school in the state. Using an AI-driven tool called Pinpoint, Sahan’s data reporter extracted and tabulated data from thousands of pages of documents. A powerful insight followed: Sahan discovered that more than $132 million in taxpayer funds was spent on outside contracts. Meanwhile, the lack of state oversight allowed schools to pay family members or departing school staff. The outlet published a series of pieces on the findings and explainers on how they conducted the investigation.

Looking ahead

As we move forward, we’re committed to sharing lessons and insights gathered from our experimentation with artificial intelligence. We plan to publish detailed case studies from our portfolio, participate in conferences across the field, and form partnerships with other organizations working at the intersection of local news and artificial intelligence. By doing so, we aim to foster a collaborative environment where best practices and innovative solutions can be shared and scaled across the local news field. 

Michael Ouimette is chief investment officer at the American Journalism Project.