On September 29, Prime Minister Rob Jetten wrote in response to parliamentary questions that “no AI was used” in posts by the Prime Minister’s official social media accounts. The Volkskrant found at least five posts by the official “minpres” accounts that were indeed written by AI.
Last month, the newspaper reported that Jetten and his party D66 had used an AI model to generate hundreds of social media posts. In response to questions about this from parliament, Jetten specifically made a distinction between his “private” and “official” social media accounts. He said that he owes parliament no explanation because he does not use AI at all on his official accounts.
The Government Information Service (RVD) also told the Volkskrant that AI was used exclusively on Jetten’s private accounts, which are managed by him and the D66, so the government’s 2025 guidelines on AI usage would not apply. The guidelines state that texts from AI may not be copied verbatim and AI-generated content must be clearly labeled as such.
Using the detection model Pangram, the Volkskrant found five posts from minpres – the Dutch Prime Minister’s handle on social media – that were written by AI and not disclosed as such. The newspaper also asked two experts to assess the posts, and both identified language characteristics of AI models. The posts all share the same rhythm and use a relatively high number of filler words to convey a message. They also all start with generic phrases like “The Netherlands and Germany are neighbors” or “The world is in turmoil.”
“It is the same cadence and structure every time,” said AI researcher Marijke van Vlasselaer.
The RVD told the Volkskrant that it does not deny that AI was used for these posts, but added that it “does not share the conclusion that the cited examples were generated entirely by AI.” The service said that the input for the posts “is supplied from various sources, such as policy dossiers, meeting notes, or previously used text,” which could explain the AI-generated nature of the text within the post.
The RVD also referred the newspaper to a new AI guideline, which states that it is not necessary to disclose the use of AI for every post as long as a human has reviewed it. According to the Volkskrant, the website’s metadata shows that the guideline was created on September 17, a week after the newspaper’s first report on Jetten’s AI use.
