Germany’s political landscape is under scrutiny following revelations that several leading politicians may have used artificial intelligence to assist in composing speeches and written statements. The development comes after former Chancellor Olaf Scholz’s aides reportedly employed an AI tool modeled on his style—dubbed the "Scholzbot"—to verify speech consistency during his tenure.
The latest investigations suggest that AI usage may be more widespread among national leaders than previously known. Mario Voigt, the chief minister of the eastern state of Thuringia, was accused of having large portions of his speeches, including two commemorations of Holocaust victims, generated by AI. Similarly, an op-ed by Karsten Wildberger, the federal minister for digital affairs, was withdrawn following allegations of AI-assisted writing.
An analysis conducted by a national newspaper applied the Pangram AI-detection software—previously employed in the cases of Voigt and Wildberger—to 92 recent speeches and articles from 11 prominent German politicians. Pangram, recognized by independent researchers as a relatively reliable method for distinguishing human-written text from material partially or fully generated by AI, flagged about 9 percent of Bundestag speeches as entirely AI-produced.
The majority of the texts reviewed were classified as entirely human-authored. Nonetheless, several politicians showed signs of partial AI involvement. A speech delivered by President Frank-Walter Steinmeier on June 13 at a national youth debating competition was identified as containing approximately 34 percent AI-generated content, particularly in a segment discussing social media’s impact on democracy. The president’s office refuted claims that AI contributed to the speechwriting, clarifying that AI tools were used solely as research aids.
Julia Klöckner, president of the Bundestag and parliamentary speaker, had three speeches flagged with partial AI composition. Notably, her April 13 lecture at Magdeburg cathedral was assessed to be 45 percent AI-generated. Officials from Klöckner’s office denied regular AI use in speech preparation and suggested that conventional rhetorical devices common in her addresses might have triggered false positives in AI detection.
Two cabinet ministers also showed evidence of AI-assisted writing. Reem Alabali Radovan, the Social Democratic minister for international development, had two speeches marked as roughly 19 percent AI-generated. Her office acknowledged collaboration among staff and conceded that AI-generated passages may have inadvertently been included in the final versions. Meanwhile, an op-ed by Katherina Reiche, the economics and energy minister, published in the Frankfurter Allgemeine Zeitung, was evaluated as 71 percent AI-generated; only the introduction appeared free of AI influence. Reiche’s spokeswoman confirmed that AI is a routine component of the speechwriting process for her team.
These findings have stirred debate about the role of artificial intelligence in political communication, raising questions about transparency and authenticity in public discourse. Some officials maintain that AI serves as a research or drafting tool supporting human creativity, while critics caution against the potential dilution of individual accountability and originality in political expression.