The Lenfest Institute for Journalism announced an expansion of its Lenfest AI Collaborative and Fellowship Program on September 28, 2026, backed by increased support from OpenAI. Launched in 2024 with AI engineering fellows at 11 major American news organizations, the initiative has grown into what the announcement describes as the largest AI fellowship program in American journalism.
OpenAI is committing a new $5 million, alongside up to $5 million in software credits and engineering support, doubling its previous support for the program. The expansion also forms part of OpenAI’s broader work with journalism organizations.
The fellowship embeds full-time AI technologists in news organizations to develop practical applications, support business sustainability, and establish responsible adoption policies. After two years of pilot work, the Lenfest Institute plans to invite another group of news organizations to participate.
Fellows collaborate with newsroom leaders, reporters, product teams, revenue staff, and executives. Their assignments combine organization-specific engineering projects with transparent, ethical policies intended to protect editorial independence, human judgment, expertise, and audience trust. Several fellows are expected to become permanent employees at participating organizations.
Across the original 11 organizations, projects have supported audience engagement, news products, investigative research, advertising, reader revenue, and efforts to extend reporting impact and community reach. Locally hired engineers became part of their host teams, learning newsroom workflows and business needs rather than approaching adoption solely as a technical exercise.
Lenfest Institute Executive Director and CEO Jim Friedlich emphasized that trust, internal collaboration, clear policies, and communication were central lessons from the first two years. He described the expansion as an opportunity for more publishers to experiment while keeping editorial standards and human judgment central. Tom Rubin, OpenAI’s Chief of Intellectual Property and Content, highlighted the practical uses journalists had developed and the importance of local reporting to community connections.
Applications developed during the pilot
When the fellowship began, generative AI had seen limited testing within local news organizations. Participating teams subsequently developed applications spanning both editorial work and business operations.
At The Philadelphia Inquirer, fellows supported Dewey, which helps journalists search decades of archived reporting. They also helped develop Scrape, a monitoring tool that replaced a reporting task requiring roughly 15 hours per week with a daily digest of potential regional story leads.
Chicago Public Media used AI-assisted translation workflows to produce time-sensitive Spanish-language coverage much faster than before. It also used AI to transcribe audio archives documenting Chicago news and culture that had remained in storage for decades.
Other organizations pursued advertising prospecting, donor modeling, personalized audience experiences, subscription growth, public-meeting monitoring, transitions from print to digital production, and operational efficiency. Together, these projects illustrated uses beyond reporting productivity while retaining journalists’ responsibility for independent, trusted coverage.
Work also spread between participating newsrooms. Fellows exchanged code, adapted ideas, compared implementation strategies, and reused approaches developed by peers, allowing individual projects to inform solutions elsewhere.
Lessons from two years of embedded engineering
- Build trust through participation. Locally hired engineers spent two years inside organizations, learning editorial routines, culture, and user needs. Listening to staff and developing tools alongside them helped fellows become trusted colleagues. As those relationships strengthened, journalists increasingly brought forward their own ideas and incorporated tools into everyday work.
- Define the need before choosing the technology. The strongest projects began with a specific operational challenge, not a requirement to use AI. Teams relied on user collaboration, limited experiments, continuous evaluation, and evidence that a solution improved the work.
- Share progress across organizations. The Baltimore Banner drew on The Philadelphia Inquirer’s work when developing its own news discovery tools. Similar exchanges of code, product concepts, technical methods, and implementation lessons helped participating organizations progress together.
Expanding participation and reusable resources
The next phase will retain embedded fellows while broadening participation to a wider range of news organizations and markets. Its focus remains helping publishers adapt to the needs of modern audiences.
The Lenfest Institute also plans to strengthen its technical capacity to turn promising fellowship projects into reusable tools, frameworks, plugins, implementation guides, playbooks, and other technical resources. These materials are intended to benefit hundreds of news organizations beyond the participating cohort.
Implementation support is meant to make the pilot’s informal cross-newsroom exchanges more durable and easier to scale. A workflow or technical advance developed for one publisher could then become shared infrastructure for others.
A broader model for local journalism
The expanded program combines embedded engineering expertise with common tools, technical infrastructure, and peer learning. Lenfest’s objective is to extend innovations from a small group of newsrooms to hundreds more.
The renewed partnership centers on giving news organizations the expertise and collaborative support needed to apply AI safely. Its stated goals are stronger journalism, more sustainable businesses, and better service to local communities.
Related updates
OpenAI Academy’s two-year milestone — September 23, 2026.
Airbnb’s expanded access to GPT-6 Astra — September 23, 2026.
An advisory group focused on mathematics and artificial intelligence — September 21, 2026.