Swiss newsrooms use AI as an assistive tool, with journalists retaining editorial control

Swiss newsrooms are putting AI to work on routine tasks, while keeping journalists responsible for what reaches readers. Media groups use the technology to transcribe interviews, summarise material and suggest headlines. Tamedia’s head of data and AI, Nadia Kohler, describes AI primarily as a tool for newsroom work.

That approach is shared by NZZ, Ringier, CH Media and Swiss public broadcaster SBC. When AI assists with writing, journalists must check the material, add context and revise it where needed. NZZ deputy editor-in-chief Barnaby Skinner says the newspaper will not allow an editorial to be written entirely by AI.

The companies see efficiency gains as a reason to adopt the tools, but they cannot yet put a precise figure on the time or money saved. The technology is already part of reporting and editing workflows, while decisions about accuracy, context and publication remain in human hands.

Readers are more willing to read journalism produced without AI assistance than material created with it. A 2024 survey by the University of Zurich’s Research Centre for the Public Sphere and Society found that 83% of respondents would read articles written by journalists without AI support. The share fell to 54% for articles created with AI assistance and 24% for texts written entirely by AI.

The findings point to a distinction in public appetite between AI used behind the scenes and AI taking over the writing. In a separate experiment, participants who did not know who had written a text did not rate AI-generated work as worse than human-written work. After learning that AI had produced the texts, they were more inclined to continue reading them. That reaction did not translate into a desire to consume more AI-generated news in future.

The survey also found a clear demand for transparency: nine in ten respondents wanted AI-generated content to carry a clear label.

AI is already affecting newsroom staffing at CH Media, particularly in copy editing. Editor-in-chief Patrik Müller says the company’s copy-editing department has been reduced because AI can correct texts effectively. The example shows that the technology’s role extends beyond experimentation with summaries and transcription.

Recent job cuts at Tamedia and SBC came alongside statements that AI would play a larger role in newsroom operations. At SBC, which owns Swissinfo, leaders view AI as one possible way to reduce costs during a financially difficult period. Thierry Zweifel, head of SBC’s data and AI business division, has said that savings pressures are driving the organisation to consider tools that could help meet them. He also acknowledges that AI can contribute to cost reductions and, in some cases, job cuts.

The companies’ adoption of AI therefore sits alongside restructuring. They expect the tools to improve efficiency, but none can yet quantify the savings in time or money.

Editorial checks and clear disclosure are central to how Swiss newsrooms manage AI use. The Swiss Press Council recommends that news organisations explain where and how they use AI tools. In practice, such disclosures remain rare, despite the survey finding that nine in ten people want AI-generated content to be clearly labelled.

For journalists, the immediate work involves checking and shaping material before publication. That can mean verifying an AI-assisted text, adding context or changing its wording. NZZ has drawn a firm line around editorials, while Tamedia says human review is required whenever AI supports writing.

The efficiency case is still developing. Ringier strategy head Patrick Rademacher says the company has no comprehensive figures yet and is working to assess whether individual AI applications deliver results. Across the industry, media groups are testing the tools while journalists retain editorial responsibility. Their next measure of success will include not only operational gains, but whether their processes make AI’s role understandable to readers.