Ten Latin American newspapers use AI to win back the young readers who no longer come in through the front page

By Zendoric · September 9, 2026.

Ten leading newspapers in Latin America — among them El Nacional of Venezuela, O Globo of Brazil, El Tiempo of Colombia and El Universal of Mexico — began an eight-month program on August 26, 2026 to design AI-powered journalism products aimed at readers between 18 and 28. The initiative, called AI-Powered Next Gen News Sprint LatAm, is coordinated by the News Product Alliance (NPA) with backing from the Google News Initiative (GNI), Google's program for funding and training newsrooms. The rest of the participants are La Nación of Argentina, La Nación of Costa Rica, La Prensa Gráfica of El Salvador, El Comercio of Peru, El Nuevo Día of Puerto Rico and Listín Diario of the Dominican Republic, all of them part of the Grupo de Diarios América (GDA).

The figure that justifies the program is an uncomfortable one for the business: according to numbers cited by the organizers themselves, readers aged 18 to 28 today account for between 2% and 5% of subscribers or of revenue at most news organizations. This is not just a matter of editorial taste: it is the first generation that discovers the news through platforms — social networks, aggregators, apps — rather than a publication's front page, and that is not going to gradually adopt its parents' consumption habits with age, as earlier generations did.

Each participating newsroom forms a four-person team — executive sponsor, editorial lead, product and AI lead, and audience lead — that first researches its own young audience, then builds a prototype during an intensive week of design and, finally, with support from specialists, turns it into a real product in production. The starting theoretical framework is the Next Gen News 2 study, produced by FT Strategies and Northwestern University's Knight Lab. NPA had already tested this same path in early 2026 with twelve local newsrooms in the United States which, according to the organization, managed to launch new experiences for that same age bracket and build in-house capacity for AI product management.

It is worth not confusing this program with the more familiar narrative about AI and journalism, the one about replacing reporters with language models. Here artificial intelligence does not write the story: it is used as a product tool, to research consumption habits and prototype formats, with the aim of recovering a direct relationship with the reader, not of cutting newsroom costs. It is a defensive application, conceived for the survival of the business, not a promise of editorial productivity.

There is an underlying tension the program does not mention but that is worth flagging. With one hand Google funds these newspapers' attempt to recover traffic and a direct relationship with the reader, while with the other it rolls out AI-generated summaries in its search engine. Broadly speaking, media outlets and industry studies have documented that this kind of summary reduces the clicks reaching news websites. It is not unique to Google — the entire generative AI ecosystem diverts referral traffic — but it puts the company simultaneously in the role of cause of the problem and funder of the solution. That does not invalidate the program; it explains why it suits Google to have healthy media outlets still around to negotiate content and distribution with.

Our reading is that this experiment opens a new front in how AI reorders value and employment within a sector, in this case journalism: the question is not which tasks a model automates, but which profiles gain weight inside the newsroom. The program institutionalizes a role that did not exist at most of these newspapers a few years ago — the "product and AI lead" — and folds it into the core team alongside the editorial lead. It is the same logic we already see in banking, health care or law: AI does not replace an entire profession, it redistributes prestige toward those who can translate the technology into a product that is useful to the end user.

In the long run, this kind of initiative fits Zendoric's underlying thesis: the abundance AI promises is measured not only in curing diseases or multiplying resources, but also in making product design so cheap that newsrooms in small advertising markets — the Dominican Republic, El Salvador, Puerto Rico — can experiment with artificial intelligence at the same pace as a major newspaper in a global capital. Progress on this particular sprint will be public at newsproduct.org; what really matters is not whether these ten prototypes survive, but whether Latin American journalism learns, with this generation of projects, to design for a reader who is never going back to the front page.