Continuing my theme of predictions for AI in newsrooms this week, and following my post on how newsrooms are moving from low-hanging fruit to AI fluency, today I offer more insights from industry leaders.

This year is shaping up to be an important one on the topics of news formats and a renewed focus on journalism:

Going bold on liquid content

Katharina Neubert vice president/strategy and investments, Holtzbrinck Publishing Group, United States:

Younger audiences consume more media than ever, they just don’t read (as much). What’s the value of a great story if few people read it? 

We will see more newsrooms move away from static text articles toward bolder liquid content experiences. By 2026, these dynamic multimedia formats will shift from early experiments to a core capability in leading news organisations. 

What’s driving this trend? Exponential model improvements, AI-first companies building these experiences based on publishers’ content — with or without consent — and increasing AI adoption by consumers.

Instead of releasing a single story, teams will test bolder, interactive formats that align with how people prefer to consume information. I’m not referring to AI-narrated articles. I’m talking about personalised audio, interactive data journalism, gamified storytelling, conversational layers, and personalised news paths. 

Why bother? Learn what resonates with existing and new audiences, and what doesn’t, before your audience decides to fully turn to platforms and AI-first media startups for news and advice.

Focus on what we do differently

Rohit Saran, managing editor, The Times of India, and Ritvvij Parrikh, senior product leader, The Times of India:

Publishers will double down on one of two extremes or both (the large publishers could attempt both):

• High-Voice UX: Focusing on distinct narrative craft (e.g., The Atlantic) where the experience of reading is the product, not just the facts. The goal is “irreducibility” — writing where the AI summary feels like a hollow substitute.

• High-Utility UX: Standardised production formats (e.g., Axios’ Smart Brevity) optimise “Time to Value.” This also improves machine legibility and can make the publisher the preferred “structured data source” for AI models to cite. 

AI will force newsrooms to stop doing the things we should have stopped long ago

Martin Schori, director of editorial AI and innovation, Aftonbladet, Sweden:

In 2026, AI won’t just reshape journalism, it will finally kill off some of the work we should have abandoned long ago. Instead of building AI-powered rewrite tools, more newsrooms will ask a more uncomfortable question: Why are we doing rewrites at all?

And instead of relying on AI to summarise overly long stories, we might simply start writing with more clarity and intention from the start.

As AI companies continue to scrape our sites, repackage our journalism, and publish it elsewhere, news organisations will focus on the stuff that can’t be distilled into three bullet points: original reporting, distinct perspectives, sharper analysis, and stories with a human voice. 

The platforms that steal our content are, ironically, forcing us to rediscover what makes it valuable in the first place.

The result? 2026 becomes the year newsrooms stop doing what never created value and refocus on ... journalism.

The year publishers unite on AI copyright

Jovan Protic, managing director, Axel Springer, Poland: 

2026 will be the year publishers finally acknowledge that the only viable path to protecting their intellectual property in the AI era is a united front. 

After two years of fragmented experimentation, we will see the first serious, organised cross-publisher initiatives aimed at defending rights, enforcing licensing, and establishing the economic value of professional journalism in AI training ecosystems.

At the same time, I expect a major shift in the behaviour of search engines — including AI search. As their models become increasingly saturated by synthetic content, the incentives flip.

In 2026, we will see the early signs of a new “AI-spam update,” where search engines begin deprioritising content that is clearly or fully AI-generated. They will need authoritative, human-created information to maintain model quality, exactly the way Google once cracked down on link farms and SEO spam.

This will also be the year when AI-generated video — long-form explainers, documentaries, and narrative formats — moves from novelty to a mainstream production tool.

Finally, in e-commerce, 2026 will expose a misconception: the idea that agentic checkout is essential. AI companies will realise that reliable, verified product feeds matter far more than closing the transaction. This will push them toward deeper collaborations with marketplaces and retailers that provide structured, vetted data — often explicitly excluding the checkout layer.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

Header art: Adobe Stock Maya Kruchancova.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .