OpenAI x Schibsted: AI-Powered Media Workflows in 2025

A lot of media teams are discovering a frustrating truth: publishing more doesn’t automatically mean growing more. The bottleneck isn’t always ideas or even reporting—it’s the operational grind around modern digital journalism: formatting, translating, tagging, clipping, summarizing, repackaging for multiple platforms, and doing it all fast enough to matter.

That’s why the news that OpenAI is partnering with Schibsted Media Group is more than a corporate announcement. It’s a signal that the next phase of AI in media won’t be about flashy demos. It’ll be about workflows—the unglamorous systems that decide whether quality journalism can scale without burning out the people who make it.

This post is part of our “AI in Media & Entertainment” series, where we track how AI personalizes content, supports recommendation engines, automates production, and helps teams understand audience behavior. The OpenAI–Schibsted partnership sits right at the center of that theme: AI isn’t replacing the newsroom—it’s becoming the infrastructure around it.

Why OpenAI–Schibsted matters for AI in journalism

Answer first: This partnership matters because it reflects a pragmatic shift: leading media companies want AI that improves speed, consistency, and distribution—without sacrificing editorial standards.

Schibsted Media Group has been a major player in Nordic media and digital services, operating news brands and subscription businesses where trust and retention are everything. When a group like that partners with a U.S.-based AI leader, it’s usually for three concrete reasons:

1. Scale content operations without scaling headcount at the same rate. Not to cut journalists, but to reduce the “everything else” work that consumes time.
2. Improve product experiences in subscriptions. Personalization, search, and discovery are now part of “the journalism.”
3. Standardize governance. Partnerships tend to come with clearer rules around data handling, tooling access, auditing, and newsroom controls.

From a U.S. technology and digital services lens (and from a leads-focused business perspective), the bigger takeaway is this: AI adoption is moving from experiments to procurement. Media organizations are no longer asking, “Can it write?” They’re asking, “Can it fit inside our stack, comply with our policies, and measurably improve cycle time?”

What AI-powered newsroom workflows actually look like

Answer first: In 2025, the most valuable AI in media isn’t a “write my article” button—it’s a set of modular capabilities integrated into content production and distribution.

If you’ve only seen generative AI used for drafting text, you’re missing where the real ROI shows up. In practice, AI tends to land in repeatable, high-frequency tasks that sit before and after publishing.

Editorial assistance (without handing over authorship)

A well-run newsroom doesn’t need AI to invent facts. It needs AI to reduce friction. Common patterns:

• Briefing and background synthesis: Turn a folder of documents, transcripts, and past coverage into a structured brief.
• Headline and dek variants: Generate options aligned to a publication’s style guide.
• Tone and clarity checks: Rewrite for readability while preserving meaning (especially helpful for breaking news updates).
• Interview transcript shaping: Summarize long interviews into quotable sections and a narrative outline.

Here’s the stance I’ve settled on after watching teams implement this: AI should accelerate judgment, not replace it. The journalist remains responsible for facts and framing; AI speeds up the mechanics.

Multiformat distribution at scale

Distribution is where modern media teams quietly lose hours. AI helps by converting one story into multiple “native” formats:

• Article-to-push notification options tuned for urgency and length
• Article-to-social variants that match platform norms
• Article-to-audio scripts for quick explainers
• Article-to-newsletter summaries with context and links to related coverage

This matters because digital content creation now includes a half-dozen surfaces beyond the homepage. AI reduces the tax of being everywhere.

Metadata, search, and internal discoverability

Media organizations sit on years of archives that are underused because they’re hard to search meaningfully.

AI can:

• Auto-generate tags and entities (people, places, organizations)
• Improve semantic search for editors and readers
• Suggest related coverage for internal linking and reader retention

For subscription businesses, this directly impacts time-on-site and content discovery, which are closely tied to churn.

Personalization and recommendations: where AI hits revenue

Answer first: AI personalization matters because it turns content volume into subscriber value—serving the right story at the right time to the right reader.

Media personalization is often misunderstood as “filter bubbles.” The reality inside subscription media is more operational: readers want relevance, and publishers want consistency.

In 2025, AI personalization and recommendation engines typically support:

• Smarter onboarding: “Follow these topics” becomes a dynamic model of interests.
• Session-to-session continuity: Resume stories and themes, not just individual articles.
• Balanced recommendations: Mix habitual interests with editorially important coverage.
• Newsletter personalization: Segment based on behavior, not just demographics.

A practical approach I like is the three-lane model:

1. For you (behavioral relevance)
2. From editors (editorial priority)
3. Around your area (local/community relevance)

It’s simple, explainable, and keeps editors in the loop. Partnerships with AI companies are often about enabling this kind of product sophistication without building every model from scratch.

The hard part: trust, rights, and governance

Answer first: The organizations winning with AI in media treat governance as a product requirement, not a legal afterthought.

When the RSS source is blocked (as it is here), it’s tempting to fill the gap with generic claims. Don’t. The real story in most publisher–AI partnerships is the same set of decisions every serious media org is making right now.

“Can we control how the tool behaves?”

Media companies need confidence that AI outputs won’t:

• Hallucinate facts
• Mimic sensitive internal language in the wrong context
• Expose sources or embargoed information

That leads to controls like:

• Approved use-cases per role (reporter vs. editor vs. audience team)
• Logging and review workflows
• Red-team testing for risky prompts

“What happens to our content and data?”

Publishers care about:

• Training and usage boundaries (what’s retained, what’s not)
• Access controls (who can use which data)
• Retention policies aligned to compliance and newsroom policy

Even if you’re not a publisher, this is the lesson to steal: AI procurement is now a governance project. That’s true across U.S. digital services—from media to healthcare to finance.

“How do we keep humans accountable?”

A strong newsroom AI policy usually states:

• AI can assist drafting, summarizing, translating
• Humans verify facts and approve final copy
• AI outputs must be reviewable and attributable internally

One-liner worth keeping on a sticky note: Accountability can’t be automated.

What this partnership signals about U.S. AI leadership

Answer first: Partnerships like OpenAI–Schibsted show how U.S. AI firms are exporting not just models, but operating patterns for digital transformation.

The U.S. has become the center of gravity for many enterprise AI platforms, and global media groups are adopting those tools for the same reason U.S. companies are: implementation speed.

What’s different in 2025 is that the value isn’t “AI exists.” The value is:

• Integration maturity: AI embedded into CMS, DAM, analytics, and customer data platforms
• Operational metrics: cycle time per story, time-to-publish across formats, reduction in repetitive editing tasks
• Product outcomes: improved retention, better discovery, more efficient audience growth experiments

For digital services leaders, the media industry is a preview of what’s coming elsewhere: AI becomes a layer across every workflow, and the winners are the ones who operationalize it responsibly.

Practical playbook: how media teams can apply this now

Answer first: Start with high-frequency tasks, measure time saved, and introduce governance before you scale access.

If you’re leading a newsroom, a media product team, or a digital content operation, here’s a practical sequence that tends to work.

1) Pick two “boring” use cases that happen every day

Good starters:

• Summaries for newsletters
• Headline variants for A/B testing
• Tagging and related-article suggestions
• Transcript cleanup and quote extraction

Avoid starting with: fully automated article generation. It creates more risk than value.

2) Define success metrics before rollout

Use measures that are hard to argue with:

• Minutes saved per story (median, not best case)
• Reduction in time from publish to “fully distributed” (social + newsletter + push)
• Increase in internal linking or archive resurfacing
• Reader engagement lifts on recommended modules

3) Build a lightweight governance checklist

Before expanding access, require:

• Approved use-cases and banned use-cases
• A review standard for AI-assisted copy
• Clear rules on sensitive inputs (sources, legal matters, embargoes)
• A way to report failures without blame

4) Train editors, not just writers

Editors are the control point. If editors don’t trust AI-assisted workflows, adoption stalls.

In practice, training that works includes:

• Prompt patterns aligned to house style
• “Bad output” examples and how to fix them
• A shared library of newsroom-approved templates

5) Treat personalization as editorial + product

Recommendation engines shouldn’t be a black box run only by data science. The best implementations I’ve seen have:

• Editorial input into what must be surfaced
• Transparent rules for sensitive topics
• Regular audits for bias and over-personalization

Where AI-powered media goes next

The OpenAI–Schibsted partnership sits in a bigger shift: AI is becoming part of how media is made, packaged, and discovered. The “AI in Media & Entertainment” story isn’t only about content generation. It’s about building digital services that keep audiences informed and engaged—without exhausting the teams behind the scenes.

If you’re responsible for a media workflow, a content supply chain, or a subscription product, the question isn’t whether AI tools will show up. They already have. The real question is: will you implement them in a way that protects trust while improving throughput?

If you want help mapping AI use cases to measurable outcomes—cycle time, distribution efficiency, personalization quality—start by documenting one workflow end-to-end (from reporting to publish to multi-channel distribution). That’s where the fastest wins tend to be hiding.

What part of your content operation still feels oddly manual for 2025—and what would it take to automate it without giving up editorial control?