German media AI scandal rocks two major newspapers

Tagesspiegel said it removed the pieces because Casdorff violated internal editorial guidelines. The newspaper stressed that journalistic responsibility, factual judgment, analytical classification, and the framing of a story must still belong to the author. Not the software.

That distinction matters more than it may look at first glance. Opinion pages depend on a reader’s belief that the author arrived at a view through human judgment, not through an undisclosed prompt and a polished machine draft. Once that trust slips, the damage is hard to reverse.

Why the case has rattled editors

Media researcher Vera Katzenberger of Leipzig University said Casdorff’s case goes well beyond using AI to brainstorm headlines or tidy up a draft. “This is not about help with brainstorming or research, this is about the core of journalistic work,” Katzenberger told DW.

She argued that opinion journalism plays a central role in democratic debate because it helps readers make sense of a complex world and shape their own positions. If a column is generated or heavily shaped by AI without clear disclosure, the public is not just reading a text. It is being nudged by a process it cannot see.

Katzenberger said AI has no values, no political stance, and no sense of responsibility. That is the uncomfortable truth for editors trying to adopt new tools without surrendering control. A machine can produce fluent prose. It cannot carry the consequences of a mistaken argument, a biased framing, or a misleading conclusion.

The episode also forced newsrooms to confront a practical problem. AI use is spreading through journalism in small ways first: idea generation, summarizing documents, polishing wording, checking grammar. Where does assistance end and authorship begin? That line is still blurry, and many outlets have not settled on a shared standard.

FAZ, Döpfner, and the labels debate

The Tagesspiegel case was not isolated. A few days earlier, Frankfurter Allgemeine Zeitung, or FAZ, withdrew an opinion piece by Thuringia’s premier, Mario Voigt, after it found the text had been produced with AI assistance. FAZ said it only learned about the issue after publication.

Katzenberger said the responsibility runs in two directions. Writers must disclose their use of AI honestly. Newsrooms, though, cannot simply rely on that disclosure alone. They need internal checks that reflect how AI has entered daily editorial work. Otherwise, the rules exist on paper only.

She also raised the question that many editors now dread: what level of AI help is still acceptable, and when does a text need a label? How much human contribution is enough? Different newsrooms answer those questions differently, and that inconsistency is part of the problem.

The debate drew a sharp response from Mathias Döpfner, the head of Axel Springer. According to DW, he criticized FAZ’s removal of the opinion piece and used AI prompts to build a polemical attack on the paper in an article published under his byline. He accused FAZ of rejecting modern technology and described it as “a desperate attempt by the horse-cart lobby to ban cars.”

The language was blunt. It also showed how AI can be used to intensify rhetoric, sharpen attacks, and blur the origin of an argument. That is useful for manipulation. Not for transparency.

What it means for readers and newsrooms

The German Press Council, the country’s self-regulatory body for print and online media, said editorial responsibility remains with the newsroom regardless of how content is produced. “This responsibility also applies to artificially generated content,” the council said.

Still, the council stopped short of saying every AI-assisted text must automatically carry a label. In its view, ethical review should focus on content and process, not just on whether a machine helped produce the words. Even so, the council said sanctions can apply when there is a serious breach of due care and truthfulness.

That leaves newsrooms with a harder job. AI tools are already embedded in many everyday workflows, from transcription to editing support. Readers may not notice the difference at first. But if a columnist or reporter uses AI for a text that appears fully human, the trust gap opens quickly. And readers feel it immediately.

For consumers of news, the practical takeaway is plain. Check whether a publication has a clear AI policy. Check whether authorship is transparent. Check whether corrections are swift when rules are broken. In a field built on credibility, disclosure is no longer a side issue. It is the test.

Katzenberger said regular training is now essential, especially for younger journalists who may treat AI as a standard tool before they fully understand the ethical lines around it. Newsrooms that delay those rules, she warned, tend to make them only after a reputational hit has already landed. At Tagesspiegel, that hit has already arrived, and the next round of newsroom decisions may decide whether this remains a scandal about one paper or becomes a template for how German media handles AI from here on.