AI Scandal in German Media Sparks Credibility Debate

JAKARTA — The AI scandal in German media has sparked a fierce debate over the limits of artificial intelligence in the newsroom, after two major outlets removed articles written with AI assistance without clear disclosure. The cases pulled in Tagesspiegel and Frankfurter Allgemeine Zeitung, then raised a simple but crucial question: when can AI help, and when has it gone too far into the core of journalism?

The first case came from Tagesspiegel, the Berlin-based newspaper. The newsroom removed several opinion pieces by Stephan-Andreas Casdorff, a former publisher and former editor-in-chief, after it emerged that he had used AI to draft opinion text. Around the same time, FAZ also took down a guest opinion by Mario Voigt, the prime minister of the state of Thuringia, after it later became known that the text had been created with AI assistance.

The Tagesspiegel and FAZ cases reopen old wounds

Tagesspiegel explained that AI is indeed used as a support tool in the newsroom, but it must not take over the core of editorial work. In its statement, the outlet said journalistic judgment, information selection, analysis, and writing style must remain the responsibility of the author. The principle sounds firm. But the Casdorff case showed that the boundary in practice is not always as clear as it looks in internal guidelines.

Casdorff himself admitted fault. He said his actions had damaged the publication’s reputation and his own. “I made a major mistake, damaging the reputation of the publication and my own reputation,” Casdorff said in an apology. “I used AI in the text. I should have explained that and therefore should not have let it be published.”

Tagesspiegel then temporarily withdrew those articles while awaiting a detailed review. FAZ took a similar step on the opinion piece linked to Mario Voigt. The paper said it only learned about the AI use after publication.

Why the case is seen as dangerous

For many observers, the problem is not simply the writing tool. Vera Katzenberger, a media researcher at the University of Leipzig, said the Casdorff case is serious because it strikes at the heart of public trust in the press. “This is not about help with brainstorming or research, this is about the core of journalistic work,” Katzenberger told DW.

She also pointed to the role of opinion in democratic debate. Opinion columns help guide readers through an increasingly complicated world. Readers use those texts to shape their own views. If an opinion is produced by AI without clear disclosure, the process of forming public opinion is also disrupted. “AI has no values, no political stance, no sense of responsibility,” she said.

At that point, public concern makes sense. Readers see the author’s name and trust the experience, perspective, and responsibility attached to it. Once it turns out the text was assembled by a machine, that trust cracks. And once it cracks, it is hard to repair quickly.

Rules are still uneven, and editors are being urged to pay closer attention

Katzenberger sees one positive lesson in the episode: newsrooms are still taking their internal policies seriously. Breaches like this, she said, must carry consequences. She also stressed that authors submitting drafts or guest opinions should be open about whether and to what extent AI was used.

On the other hand, editors cannot rely only on the author’s word. That is why, according to Katzenberger, verification procedures need to change as well. Editorial teams need clear rules: what kind of AI assistance is allowed, when content must be labeled, and how much personal contribution is expected from human writers.

The problem is that practice on the ground has moved faster than policy. For many journalists, AI now feels like a search engine or spell-check tool: part of everyday work. The gap between acceptable assistance and AI authorship that must be disclosed is getting harder to see. That is where the risk begins.

Press Council approach and criticism from Axel Springer

Germany’s Press Council, the self-regulatory body for print and online media, said full responsibility for all editorial reporting remains with the newsroom, regardless of how the text was produced. That rule also applies to AI-generated content. Still, the council does not see a special labeling requirement for AI text as something that should be imposed.

Its view is that ethics reviews focus on content and accuracy, not on who or what made it. Even so, the council leaves room for cases that could be seen as serious breaches of care and truthfulness.

Mathias Döpfner, CEO of the influential Axel Springer media group, sharply criticized FAZ’s decision to pull the AI-assisted opinion piece. In his comment, he said FAZ was rejecting modern technology and compared the stance to a desperate effort “by the livery stable lobby to ban the car.” His remark showed one thing clearly: the debate over AI in media is not only about ethics, but also about the future direction of the industry.

Axel Springer itself is no minor player in this debate. In March, the company’s Business Insider site was criticized publicly for publishing an AI-written report about a mother of a toddler working from home, while naming a human author on the article. That piece was later removed. A string of cases like this has only deepened public doubt.

What is at stake: reader trust

For readers, the impact is obvious. When a newsroom is not transparent, readers lose the ability to judge whether an opinion comes from experience, research, or a machine’s output. In journalism, trust does not come from site design or publishing speed. Trust is built on honesty.

Katzenberger called routine training and open discussion the most sensible path. She urged students and young journalists to see AI as a tool, not a replacement. That view feels relevant in many newsrooms, including in Indonesia, where AI tools are also becoming familiar for research, transcription, and first-stage editing.

But the core message stays the same. Machines may help. The person in charge remains human. If that line blurs, what weakens is not just one article, but the media’s own legitimacy.

“We need to clearly define what forms of AI assistance are still allowed, when content must be labeled, and how much personal contribution must be present,” Katzenberger said. “If that is not regulated, the boundary between legitimate assistance and AI authorship will keep getting blurrier.”

Quick summary

• Two major German media outlets removed articles after undisclosed AI use triggered criticism.

• The cases are seen as dangerous because they affect public trust and the core of journalism.

• Newsrooms are now being urged to set clearer rules on when AI may be used and when it must be disclosed.