It’s that time of the year when we are all thinking about the year ahead and what it might bring.

In that spirit, this blog focuses on predictions for AI in newsrooms. I spoke to insightful AI practitioners and strategists from across the world who have kindly contributed their thoughts.

Start your (AI) engines

Lyn-Yi Chung, deputy chief editor, MediaCorp, Singapore: 

A few clear themes have emerged for publishers hoping to reengage an increasingly slippery audience: Create genuinely accessible content, lean into creator-style video, and rebuild relationships, community, and daily habits.

If there is one arena that is undeniably ripe for AI intervention, it is video. Production remains time-consuming, resource-heavy and full of friction, especially when it comes to vertical formats where “speed to publish” is everything. AI can and should collapse this workflow from hours to minutes.

It’s time for newsrooms, news agencies, and solution providers to raise their ambition and deliver tools that match the pace and expectations of current audiences.

But workflow acceleration is only the entry point. The next frontier is using AI to architect interactivity and community at scale. Newsrooms need to stop thinking about AI purely as an efficiency engine and start thinking about it as a relationship engine.

My prediction for 2026: Newsrooms will finally move beyond the low-hanging fruitof AI — optimising and reversioning text, deploying chatbots as glorified search, and other incremental wins. The real leap forward comes when they focus not on what AI can easily do but on what AI can meaningfully do.

The top strategic priority for every newsroom will be this: Ensure your data infrastructure and content are fully indexable, searchable, and actionable. Journalists and editors often resist meta tagging because it feels like work done for machines, not humans.

But in an environment where discovery is algorithmic, attention is scarce, consumption habits are shifting, and news fatigue is a constant undertow, structured data is no longer optional. It’s the foundation for visibility, personalisation, and every future AI capability that matters.

AI fluency becomes editorial capital

David Cohn, senior director, Alpha Group, Advance Local, United States:

There’s a long history within newsrooms to silo certain technical skills. CAR (Computer-Assisted Reporting) was once a specialty of a few investigative reporters who could navigate databases and spreadsheets. By blending editorial instincts with technical skills, they could do a kind of reporting that felt out of reach for most. 

In 2026, the boundary between “generalist editor” and “technical specialist” will begin to erode. While AI continues to support table stakes tasks such as transcription and grammar checks, more will step into AI not as passive users but as hands-on architects of workflow change. 

For the past two years, AI experimentation has lived mostly with product and R&D teams, creating a cultural gap between the people who built tools and the people who set editorial standards. That gap narrows rapidly as editors learn prompting, evaluate model behaviour, understand retrieval pipelines, and participate directly in tool design.

AI governance can’t be outsourced, and the newsroom’s values can’t be inferred. If it’s true that in 2017 the CMS was your first editor, in 2026 AI is your editorial engine room. It shapes not just what you publish but how the newsroom thinks, organises, and even governs its work. 

The role of the editor evolves from steward of storytelling to steward of systems. AI literacy becomes essential for leadership, which is part editorial judgment and part systems thinking. In 2026, the newsroom doesn’t “adopt” AI. Instead, newsrooms expand what counts as editorial expertise.

Human roles — and skills — change dramatically

Ana Jakimovska, head of AI strategy, Mediahuis, Belgium: 

As AI becomes ubiquitous in news organisations, here’s what’s going to change:

• Research becomes radically faster. Reporters use AI as the first filter: summarising documents, surfacing what’s new, cross-referencing facts. The background work that takes an hour today takes five minutes tomorrow. NotebookLM will summarise hundreds of pages in minutes.

• Copy editing shifts from grammar to judgment. Tools handle spelling, structure, clarity. Editors focus on intent, tone, defensibility of claims, and whether the story delivers the brand’s signature value.

• Photo desks split into curation and controlled generation. Curating real images stays crucial. But another large task emerges: generating safe, brand-consistent visuals for every platform and format. “Give me three variants in our house style, horizontal, vertical, reels.”

• Data literacy becomes default. Not heavy coding, but every reporter will have tools to explore datasets, detect trends, run comparisons, and generate charts they then interpret and verify journalistically.

• Transcription, translation, and cleanup disappear as tasks. Everything becomes instant and reliable. 

• Headline writing becomes almost entirely a machinated task. Not to kill creativity but because the system can instantly generate variants tuned to brand voice, SEO, subscription strategy, and target audiences. Humans steer; AI produces volume.

• Fact checking becomes more meta. Less: “Is this fact true?” More: “What sources is the AI relying on, and are they valid?”  Journalists will increasingly verify the lineage of information rather than singular claims.

• Production routines fall away. Teasers, newsletter intros, social captions, push proposals — these become assisted or automated flows. A human still decides, but the repetitive production work fades.

• Adding enrichment becomes easy and fast. The tools are already in place to generate videos and audio bulletins within minutes.

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

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