Lenfest expands AI fellowships with OpenAI support

The expansion combines committed funding with potential in-kind support. Its aim is to help more local newsrooms adopt tools developed inside the fellowship.

The Lenfest Institute for Journalism, a nonprofit supporting news sustainability, announced an expanded AI fellowship on September 28 with $5 million from OpenAI, the AI developer, plus additional in-kind support.

The program places technical staff inside news organizations. Its next phase aims to share useful tools more widely. Newsrooms will need to decide which tools improve their work and how to sustain them after the initial support.

Why it matters: A newsroom may need engineering time and implementation help as much as access to a model. The expansion could support that work, but a funded experiment should still be judged by editorial usefulness, operating cost and the ability to maintain it.

The announced package separates the $5 million commitment from up to $5 million in software credits and engineering support. The second amount is a ceiling on additional resources. It should not be described as another unconditional cash grant or as money already received by participating newsrooms.

Lenfest says the next phase will retain embedded fellows, broaden participation and turn promising projects into reusable resources. The ambition is wider adoption beyond the original organizations. That is a program plan, not evidence that hundreds of newsrooms have already deployed the resulting tools.

The program began in 2024 with support from OpenAI and Microsoft, another technology provider. Lenfest’s program page describes an initial group of 11 news organizations with two-year fellows and shared code and implementation lessons. That background explains why this announcement is an expansion rather than a new fellowship starting from nothing.

For a small newsroom, the value of shared work depends on how much adaptation remains. A useful evaluation would ask whether a tool can be installed with the newsroom’s existing records, whether staff understand its limitations and who will fix it when the underlying system changes.

The distinction between a prototype and a maintained tool matters here. A demonstration can show that an idea is possible. A production decision requires a named owner, a repeatable review process and a budget for continued operation. Those are proposed adoption criteria, not a claim that every existing fellowship project lacks them.

Judge the work after the subsidy ends

Credits reduce the initial expense of experimentation, but they are different from a permanent operating budget. Before adopting a tool, a newsroom could estimate its ordinary usage cost after the credits end. That exercise would help distinguish a sustainable improvement from a project that works only while an external subsidy covers it.

Editorial quality needs an equally concrete test. For an archive-search assistant, evaluators might compare returned source passages with the original reporting and record incorrect or missing references. A fluent summary should not be counted as a successful result merely because it is easy to read.

For a tool that proposes story leads, the newsroom could count useful leads that survive a reporter’s checks. Time saved before verification should be considered together with time spent correcting false leads. The purpose of such a trial would be to measure the actual workflow rather than only the generation stage.

Vendor support also raises an institutional question distinct from technical quality. A newsroom should preserve its editorial authority when evaluating technology supplied by a company it may cover. Clear disclosure and independent editorial decisions would let readers understand the relationship without assuming that a grant determines coverage.

Lenfest presents its pilot as successful. That assessment is attributable to the program operator; it is not an independent finding that the program caused better business results. A stronger evaluation would report costs, sustained use and documented outcomes, including projects that did not prove useful.

The next concrete developments will be the new participation details and the reusable tools and support materials that emerge from the expansion. The funding creates an opportunity to test and maintain newsroom technology. Whether it strengthens local journalism will depend on the work those newsrooms can verify and continue using.

Verification

Glossary candidates: in-kind support — resources supplied instead of cash; embedded fellow — a specialist working within a host organization; software credits — an allowance against a provider’s charges.

Cold-reader sentence: Lenfest expanded its newsroom AI fellowship with $5 million committed by OpenAI and up to $5 million more in credits and engineering support.