Swiss media companies are expanding their use of artificial intelligence, but none can say how much time or money the technology saves. A second gap has also appeared: readers want AI-generated content labelled, yet newsrooms rarely do it.
The findings come from a report by SRF journalist Jürg Tschirren, published on November 5, 2026. It looked at how major Swiss media groups use AI in daily news work.
Newsrooms are using AI to transcribe interviews, write summaries and suggest headlines. Nadia Kohler, head of the data and AI unit at Tamedia, describes AI mainly as a tool.
Tamedia says AI should not produce journalistic texts without oversight. When AI is used in writing, humans must check the content, add context and change it if needed.
NZZ, Ringier, CH Media and Swiss public broadcaster SBC take a similar approach. Barnaby Skinner, deputy editor-in-chief of NZZ, says the paper will never let AI write an editorial on its own.
Savings Remain Unmeasured
All of these companies expect AI to make their work more efficient. Still, none can put a figure on how much time or money it saves today.
Patrick Rademacher, who leads strategy at Ringier, says no complete figures exist yet. The company is trying to learn whether each AI application actually delivers efficiency gains.
A study from Denmark points to modest results. Researchers surveyed 25,000 workers, including journalists, and found that those using AI chatbots saved about 3% of their working time on average.
Even without clear numbers, AI has already affected some jobs. CH Media editor-in-chief Patrik Müller says the company’s copy editing department has shrunk because AI can now correct texts well.
Recent layoffs at Tamedia and SBC, Swissinfo’s parent company, came with statements that AI would play a larger role in newsrooms. At SBC, AI is seen as a possible way to cut costs during hard financial times.
Thierry Zweifel, who heads SBC’s data and AI business division, says AI is not what is driving the savings effort. He says the need to save money is pushing the organisation to look at helpful tools, and that in some cases this can lead to job cuts.
Readers Want Labels That Rarely Appear
Public views on AI in journalism remain cautious. 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 help.
That share falls to 54% for articles made with AI support. Only 24% said they would read a text written fully by AI.
The doubts do not seem to come from quality. In a University of Zurich experiment, participants did not rate AI-generated texts as worse than human-written ones when they did not know who wrote them.
When participants later learned AI had written the texts, they were even more willing to keep reading them. However, that did not make them want to read more AI-generated news in the future.
Nine out of ten survey respondents want AI-generated content to be clearly labelled. The Swiss Press Council also recommends that newsrooms disclose where and how they use AI tools.
In practice, those disclosures are still rare. Most media organisations do not say whether a human reviewed and approved content, no matter how much AI was involved in making it.
