Two German media outlets were compelled to retract articles that employed undisclosed artificial intelligence, raising concerns about the potential erosion of credibility within German journalism.
In a statement released over the weekend, the Berlin-based newspaper Tagesspiegel addressed this controversy by stating, "For our newsroom, AI serves as a tool to streamline and enhance specific editorial tasks. It is not, however, intended to replace the core functions of our work." The editors aimed to mitigate the fallout stemming from the revelation that one of the newspaper’s notable political commentators, Stephan-Andreas Casdorff, had utilized AI to develop opinion pieces.
Casdorff, a former publisher and editor-in-chief aged 67, acknowledged his significant error, admitting, "I have made a substantial mistake that has tarnished the reputation of the publication and my own. I sincerely apologize. I used AI in the articles and should have clarified that; they should never have been published."
In response to the controversy, the editorial team at Tagesspiegel removed several of Casdorff's AI-generated articles from their online platform, stating, "We have decided to temporarily take the affected texts offline until a thorough review is completed."
This incident has intensified an ongoing and often heated discussion regarding artificial intelligence’s role in journalism. Just days prior, it was disclosed that an opinion piece by Mario Voigt, the state premier of Thuringia published in the Frankfurter Allgemeine Zeitung (FAZ), was also AI-assisted, with the newspaper only learning of this after its publication.
Vera Katzenberger, a media researcher from Leipzig University, characterized the Casdorff situation as particularly troubling due to its potential to undermine public trust in journalism. She emphasized that this issue extends beyond merely using AI for brainstorming or research, placing it at the heart of journalistic integrity. "Readers support newspapers for the insights and expertise of specific authors. If AI generates opinion pieces without any disclosure, it could be perceived as misleading," she noted.
Katzenberger expressed concerns about AI-generated editorial content disrupting the democratic discourse. "Opinion pieces guide us through a complicated world and assist in forming our views. AI's involvement in crafting these narratives poses risks, as it lacks values, political affiliations, or accountability," she stated. Nonetheless, she also viewed the case positively in that it reflects a commitment to upholding editorial standards, along with the recognition that violations of these principles can lead to serious consequences.
The Tagesspiegel editors confirmed that the removal of Casdorff's pieces was a necessary step due to his breach of established editorial guidelines that all staff must adhere to. "The authors must bear the responsibility for journalistic judgment, analytical assessment, and the crafting of the text," they asserted.
Similarly, the FAZ has taken down the AI-assisted opinion piece by Voigt. Katzenberger holds that any contributor submitting an article should clearly disclose their use of AI while asserting that editors cannot solely rely on the claims of authors. She stressed the need for editorial teams to update their review processes to accommodate the evolving landscape of AI in journalism. "What AI support is permissible? When should AI-generated content be labeled? What amount of personal input is expected?" are critical questions that must be addressed, she noted.
Mathias Döpfner, CEO of the prominent media company Axel Springer, expressed his dissent regarding the FAZ’s choice to retract Voigt's AI-generated opinion piece. He mentioned that he had prompted AI to critique the FAZ, with the resulting text being published under his name and accusing the outlet of resisting technological progress, likening it to a "desperate attempt by the stagecoach lobby to inhibit the automobile."
The German Press Council, which oversees the country's print and online media, maintains that editorial responsibility rests entirely with the newsrooms, encompassing content produced through AI. However, the Council deems a labeling requirement for AI-generated articles as unnecessary, positing that the ethical evaluation of complaints remains unaffected by the creation process or tools used.
Earlier this year, the business news site Business Insider, owned by Axel Springer, faced public censure for publishing an AI-generated story regarding a working mother, misattributed to a specific author. This report was later withdrawn.
Katzenberger articulates the urgent need for proactive measures in light of blatant infringements of editorial standards. For many journalists, integrating AI has become as routine as utilizing search engines or spellcheckers, blurring the lines between valid support and AI authorship. She believes that through regular training and open discussions concerning ambiguous uses of AI, the situation can improve. Her advice to students is to view AI as a facilitator rather than a substitute for their own journalistic insight. "There is always the risk that relying on AI could hinder their professional growth," she warned.
Moreover, she highlighted the importance of media outlets transparently addressing mistakes. "Trust isn't built or destroyed by a single incident," she clarified. Refusing to engage with AI altogether would be unrealistic, as Katzenger stated, "AI is here to stay."
