Newsrooms don’t fail because they lack content; they fail when judgment cannot keep up.

In the age of AI, production has been optimised, accelerated, and automated; editorial decision-making has not. It remains fragmented, experience-bound, and increasingly difficult to sustain as senior expertise thins across organisations. This is not a Taiwan-specific issue, but a structural challenge facing newsrooms worldwide.

At United Daily News Group, we set out to solve this problem — not by automating decisions but by operationalising them.

Providing judgment at scale

Publish X is not a content tool. It is a decision infrastructure embedded at the newsroom’s central desk — where editorial priorities are set, resources are allocated, and accountability resides.

At its core is the democratisation of editorial expertise. Traditionally, judgment has been concentrated in the experience of senior editors — valuable, but difficult to transfer and nearly impossible to scale. Publish X transforms this model by converting decades of editorial experience into structured, shared decision logic.

Through iterative interviews and real-world validation, tacit knowledge was translated into multi-layered frameworks. Editorial processes — from reporting experience, judgment, and prediction to outcomes — were mapped into eight activation flows and 41 reporting scenarios, enabling structured, actionable story planning.

What was once individual intuition becomes an organisational capability — accessible, testable, and continuously refined.

This shift redistributes expertise across the newsroom. Editorial judgment becomes a shared resource embedded in daily workflows. In fast-moving stories, editors-in-chief and cross-functional teams can rapidly align on angles, coordinate coverage, and execute without duplication or blind spots.

Collaboration becomes a system-supported process.

Three tiers, one mission

The framework operates across three layers: human judgment remains the final authority; structured logic provides alignment; and AI acts strictly as a support layer.

In Publish X, AI does not make decisions — and is not allowed to. Every decision remains human-owned, with mandatory approval embedded at critical points. Journalists engage with AI critically, evaluating outputs in terms of context, ethics, privacy, and bias.

One of the most striking shifts is in the timing of decisions. Publish X can generate a structured story plan within two minutes. This does not replace human decision-making; it advances it. Senior editors are brought into the decision process earlier, with more time to assess blind spots, challenge assumptions, and expand coverage perspectives.

The impact has been both operational and cultural. Teams align faster, while judgment becomes institutionalised across editorial, data, and engineering teams. This has translated into sustained growth in reach, engagement, and content performance — driven by better, more consistent decisions.

The challenge facing journalism is not whether AI can produce more; it has already proven it can. The real question is whether news organisations can preserve the quality and accountability of the decisions behind that output.

Publish X offers one answer: scale judgment, without surrendering it.

In this system, AI expands what is visible, but not what is decided. The future of journalism will not be defined by how much we automate but by what we choose to keep human. 

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