We won’t get fooled again (we hope)

When I first met Gerry Lenfest, a pragmatic but visionary cable TV entrepreneur, I had to remind myself that this soft-spoken 85-year-old had helped usher in two of the great disruptions in the history of news and media: the advent of multi-channel television and ubiquitous broadband access to American homes.

The Lenfest Institute for Journalism was founded a decade ago to raise and deploy money and advice to help local journalism survive and thrive. Gerry’s original name for his Institute was prescient: “The Institute for Journalism in New Media.” I and others felt that was too much of a mouthful. I was dispatched to explain that the phrase “new media” was old-fashioned, a creature of the 1990s and no longer new. I was wrong.

The concept was — and remains — that journalism must always be new, and that independent local news should be positioned to thrive however technology and consumer habits evolve. And that the future of our business might be enabled rather than hobbled by new media. That notion will be profoundly tested in 2026 and beyond. It was in this spirit, two years ago, that we approached OpenAI and Microsoft with a proposal that they enable us to invest at scale in the use of AI to help sustain and advance the business of great journalism. We found them receptive, supportive, and eager to work together.

Each company agreed to grant the Lenfest Institute $5 million, divided equally between cash and software or cloud-storage credits. In October 2024, we launched the largest artificial intelligence fellowship program in American journalism.

All of our partners’ hard and soft dollars have been regranted to smart, deserving, and aggressive news organizations. We are working with ten newsrooms of significant size. Each hired a full-time AI engineering fellow for two years:

The Philadelphia Inquirer has built an advanced search tool to leverage its 195-year news and photo archive. Newsday is scouring large public data sets for news insights and business value. Chicago Public Media is translating stories into Spanish and expanding its potential audience sharply. The Baltimore Banner, The Seattle Times, The Dallas Morning News, and The Boston Globe are using the technology to engage prospective donors, advertisers, and subscribers. ProPublica is using AI to help direct hundreds of weekly news tips to the right reporters for further investigation. The Minnesota Star-Tribune has invented novel product formats for food, sports, events, and breaking news.

All ten organizations are eagerly sharing code on GitHub, replicating one another’s work, and spurring innovation at a greater scale.

That’s the good news.

The bad news is that beyond these 10 there are only a handful of local newsrooms with dedicated AI development staff fully exploiting the technology.

How does the local news industry’s investment in AI compare to other mission-critical American industries? In a word, pitifully.

I recently learned from a colleague that there are more AI engineers at one major U.S. health insurance provider teaching AI to read — and as often as not reject — health insurance claims than Lenfest estimates are employed across all of American local journalism.

We have been to this barbecue. The news industry’s response to the consumer Internet’s arrival in the 1990s was flat-footed, under-capitalized, and lacking in imagination, a strategic error from which we have never recovered.

Now we are about to repeat the errors of the past, almost verbatim: underinvestment in new technology and naivete about its impact; a lack of focus on new product development and the new audiences it might engage; a failure to act boldly or at scale.

So what to do?

In a phrase, meet the moment. The news industry, and local news in particular, needs to stop playing small ball with AI.

We need to move beyond a focus on AI as a means to improve internal business processes and focus much more attention—and money—on product innovation. We need larger players to lead on this effort, and smaller ones to become more nimble, replicating the most promising innovations as they emerge. We need less handwringing about the risks of AI to news, and greater effort to employ its promise to our benefit.

The 1971 song from The Who says it well: “We won’t get fooled again.” In 2026, we will see if that anthem proves a smart prediction for the American news business, or just wishful thinking.