Newsrooms use AI mostly to speed up existing work, Knight Center report finds

Newsrooms have put AI to work speeding up familiar tasks, but that activity has not added up to a fundamental transformation in journalism, a new Knight Center for the Future of News report finds. Its central tension: leaders worry about AI changing their audiences and economics, while relatively few projects tackle revenue.

Described by ASU News on October 15, The News + AI Field Report: What News Organizations Built with AI, 2022–2025 catalogs 160 projects launched by 101 U.S. news organizations. The projects span late 2022 through the end of 2025. Author Djordje Padejski, a journalism scholar and Cronkite doctoral candidate, worked with support from ASU graduate students Presley McCaskill and Tufan Neupane.

Faster production is the dominant use

Summarization, drafting, transcription, translation and workflow automation dominate the dataset. Many newsrooms are deploying production assistants to reduce repetitive tasks, with the aim of giving journalists more time for reporting. These are largely tools for work news organizations already do, rather than a different way to deliver or fund journalism.

The researchers identified six patterns across the projects and accompanying interviews with newsroom staff. Although the number of projects grew rapidly over the period, they concluded that the combined activity had not fundamentally changed how journalism is produced, distributed or funded. Their distinction is between adopting more tools and changing the operation those tools support.

Some tools reach beyond the existing workload

Not every project is simply a production shortcut. The report identifies tools that help reporters process more information than a team could handle on its own, alongside a smaller group designed to help people navigate civic life. Those uses broaden either what a newsroom can examine or the service it offers.

• Public-meeting monitors can listen across a region’s city council meetings and surface potential stories that might otherwise go unnoticed.
• Document-analysis systems, investigative research tools and podcast-monitoring projects help newsrooms work through information at a scale beyond their reporting teams.
• Civic-information experiments address property taxes, evictions, restaurant inspections and immigration questions. These public-service tools remain a minority of the projects.

Padejski argues that efficiency should create room for this kind of work, not become the endpoint of AI adoption. His proposed use for freed-up time, money and staff capacity is wider reporting, new community relationships and more useful news experiences—not simply a larger volume of content.

That matters because the biggest disruption from AI is not coming from how news gets produced, but how people find, experience and interact with news and information through AI assistants and AI search.

Djordje Padejski, report author

Business anxiety has not set the project agenda

In interviews, newsroom leaders expressed widespread concern about AI’s effects on site traffic, audience behavior and the economics of news. Yet relatively few projects in the catalog focus on revenue, subscriptions or fundraising. The gap puts the most common development priorities at a distance from some of the risks leaders themselves identify.

Adoption is also uneven. Even as the cost of developing an AI tool has fallen, the report finds activity concentrated among large national outlets, well-funded digital-native organizations and some nonprofits. Cheaper development has not, on its own, produced an evenly distributed wave of newsroom experimentation.

The researchers’ practical guidance is not to chase the most advanced technology. Projects that stood out applied AI to a specific journalistic problem rather than building for the sake of using it. The catalog is intended to help newsroom leaders direct limited time and resources by showing both what peers have built and which needs remain less explored.