Lenfest Institute expands newsroom AI fellowship with new OpenAI support

The Lenfest Institute for Journalism announced on September 28 that it is expanding its AI Collaborative and Fellowship Program for US news organizations, backed by a new $5 million commitment from OpenAI and up to $5 million in software credits and engineering support. The next phase is intended to bring more newsrooms into a program that embeds AI engineers in their operations and to make selected projects useful beyond the organizations that developed them.

The funding commitment and the additional credits and support are separate parts of the announcement. OpenAI and Lenfest described the expansion in a joint statement; Nieman Journalism Lab reported the package as $5 million in cash and up to $5 million in credits and support. The full potential value therefore depends on use of the additional support, rather than representing a $10 million cash grant.

Which newsrooms could join the Lenfest fellowship

Lenfest says it will invite a new group of news organizations into the fellowship, which began in 2024 with AI engineering fellows at 11 American organizations. Nieman Journalism Lab reported that the next cycle will consider local television news and statewide public service journalism, widening a program previously focused on metro newspapers. The joint announcement does not spell out those categories or provide application dates.

The program’s model is to place full-time technologists inside participating news organizations. According to Lenfest, they work with reporters, newsroom leaders, product teams and revenue staff on problems specific to each organization. Lenfest chief executive Jim Friedlich told Nieman Lab that fellows are hired locally and are not employees of Lenfest, Microsoft or OpenAI.

What the first fellowship produced

The Philadelphia Inquirer developed Dewey, a tool for searching archived reporting. Nieman Lab reported that engineer Kevin Hoffman built it to search hundreds of thousands of digitized articles dating to the 1970s, and that investigative, political and crime reporters have used it for historical research and verification. The joint announcement also describes Scrape, an Inquirer monitoring tool that turns what had been a roughly 15-hour weekly reporting task into a daily digest of possible story leads.

The Inquirer’s Matt Boggie told Nieman Lab that tools developed during the fellowship had processed more than 270,000 images in its photo archive. He said software credits helped pay for experimentation and deployment. Those are specific accounts of newsroom use; they do not establish a measured improvement in reporting quality or the finances of participating publishers.

Work has also moved between organizations. Nieman Lab reported that the Baltimore Banner adapted the Inquirer’s Dewey tool, while Chicago Public Media adapted a version for searching radio transcripts. The joint announcement says Chicago Public Media also used AI-assisted translation workflows for time-sensitive Spanish-language coverage and transcribed older audio archives. These examples show the kinds of tasks fellows have taken on across editorial and production work.

How the new commitment builds on the 2024 launch

The program began with support from both OpenAI and Microsoft. At its October 2024 launch, Lenfest announced that each company had committed $2.5 million in direct funding and $2.5 million in software and enterprise credits. Axios independently reported the $10 million pilot as equal contributions from the two companies, with each initially selected outlet due to receive $500,000 toward a two-year fellow.

The initial five selected organizations were Chicago Public Media, Newsday, the Minnesota Star Tribune, the Philadelphia Inquirer and the Seattle Times. Their proposed work included transcription and translation, archive search, public-data tools, content discovery and advertising analytics. Lenfest later described a fellowship operating across 11 organizations. The new announcement says OpenAI has doubled its previous support; it does not establish a renewed commitment from Microsoft for the next cycle.

What Lenfest plans to share next

Lenfest says it wants to turn the strongest fellowship projects into reusable tools, frameworks, plugins and implementation guides for other news organizations. Friedlich told Nieman Lab that the aim is to let newsrooms copy and customize useful work. The reported reuse of Dewey offers an example of that approach, but the wider distribution of future tools remains a plan rather than a demonstrated outcome.

Friedlich said in the joint announcement that the pilot had taught the institute the importance of trust, guardrails and open communication when newsrooms adopt AI. He told Nieman Lab that Lenfest would not invest in newsrooms intending to use AI chiefly to replace human reporting and writing. OpenAI executive Tom Rubin described the funding as support for people and implementation and the credits as a way to experiment. Those are the partners’ stated priorities, while application dates, the cohort size and a timetable for shared tools remain unannounced.