AI in Journalism: How Five Patterns Separate Success from Expensive Experimentation

Hi,

I have some big news today. With News Machines growing at a slow but steady pace (getting close to 1,000 unique subscribers for the English plus German version), I have decided that it’s time to launch a paid tier. I will keep the weekly case studies free for as long as I can sustain that effort, but I am launching an additional Premium Club with some extra perks such as in-depth reports that distill Gen AI case studies and into patterns and actionable insights.

I would be thrilled to welcome you as a News Machines Premium Club member. The early bird rate until December 31 is $80 per year. Starting January 1, the annual rate will be $100. 

In other news: I’m thrilled that my first-ever Nieman Journalism Lab prediction was published today.

In more other news: The holidays are getting close and News Machines will be taking a break until early January. 

Happy Holidays!

Ulrike

For my first 28-page in-depth report, "The Human-Augmented Newsroom: AI in Journalism Strategic Patterns 2025," I synthesized findings from 33 News Machines case studies and industry research into five patterns that separate successful AI deployment from expensive experimentation. 

Here is a preview: 

Three years after ChatGPT's launch, the news industry has moved from experimentation to strategic reckoning. The question is no longer whether to use AI but how to deploy it without undermining the trust and originality that make journalism valuable.

According to the Digiday State of AI survey from October 2025, 87 percent of publishers already use AI and 97 percent plan to increase investment. Yet only 1 percent have deployed AI across all workflows. That gap between intention and implementation reflects hard-won caution: The EBU News Report 2025 found issues in approximately half of AI-generated content tested across European public broadcasters.

Audiences have drawn clear boundaries. The Medill Local News Survey 2025 found 47 percent feel comfortable with human journalists using AI assistance, but only 17 percent accept content created mostly by AI, even with human oversight.

Applications that apply AI to well-defined problems with verifiable outputs show the strongest returns. The Philadelphia Inquirer, Baltimore Banner, and Wall Street Journal have each found different ways to turn historical content from storage costs into strategic infrastructure. The full report details how each approach works and what resources they require.

AI can help small teams cover beats that would otherwise go dark, but only when deployed for news gathering rather than content creation. iTromsø's DJINN tool reduced journalist research time by 94 percent. The system generates leads; reporters write stories. The report examines five local news deployments and what distinguishes the successful ones.

The industry has split between publishers licensing content to AI companies and publishers suing for damages. A third path exists: Building AI into new revenue streams. Morning Brew expects B2B revenue to surpass revenue from its flagship consumer newsletter in 2025. The report maps the emerging business model landscape and identifies which approaches suit different publisher profiles.

Features that save time build engagement. Features that replace journalism erode trust. The report covers implementations from inewsource, Deutsche Welle, Particle News, and Mediahuis, with specific metrics on what drives audience acceptance.

Ben Collins, CEO of The Onion: "AI is not funny at all. It's like the least funny technology that exists. AI picks the middle of everything. Jokes are never the middle of anything." The Onion grew to 56,000 subscribers by rejecting AI entirely. The report examines when human-created positioning works as competitive advantage and when it doesn't.

Sponsored

Connect your email, and you’ll instantly get a CRM with enriched customer insights and a platform that grows with your business.

With AI at the core, Attio lets you:

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

👉Try Attio Pro for free

Digiday found that 93 percent of AI budgets are going to infrastructure and only 6 percent to staff training. Publishers are building technical capacity faster than the human capacity to use it responsibly. Anne Lagercrantz, Director General of Swedish public broadcaster SVT: "We are increasing individual efficiency and creativity, but we're not saving any money. Right now, everything is more expensive."

The complete report includes implementation evidence from all 33 case studies, specific action items organized by publisher size (30-day, 90-day, and 6-12 month priorities), a section on fighting AI-generated misinformation when AI search tools fail basic accuracy tests, and the critical questions the industry cannot yet answer.

It's designed to save you time: You can spend five minutes for the strategic patterns and action roadmap, 30 minutes to add the learnings relevant to your organization, or two hours for the full report with links to every case study.

The full 28-page report is currently available exclusively for paid subscribers. Your subscription also supports my independent reporting on AI in journalism and helps to keep the weekly case studies a free service for everyone.