Artificial intelligence has reached the heart of the Dutch government, with Prime Minister Rob Jetten emerging as one of its enthusiastic users. An investigation by De Volkskrant found that at least 232 messages published on social media by the social-liberal prime minister and his party, D66, were written with the assistance of virtual tools such as ChatGPT.
The findings have opened a debate over the use of AI in political communication, where messages presented as personal reflections or carefully considered statements can carry particular weight. Although political speeches and statements are commonly prepared with the assistance of close advisers and staff, widespread reliance on chatbots could damage public credibility if voters believe that supposedly personal communications have instead been delegated to a machine.
Jetten defended his use of the technology, telling public broadcaster NOS: “The fact that AI is used as a tool does not mean that it is not also reviewed by human eyes.” Speaking to commercial broadcaster RTL Nieuws on Thursday, he added: “Yes, like three-quarters of Dutch people, I also use AI from time to time to improve things.”
The prime minister said he believed “we all need to think clearly about how AI can be properly integrated in 2026”. His party, D66, has supported a new European AI law requiring the mandatory identification of content generated by artificial intelligence from August 2026.
One of the messages examined by De Volkskrant appeared on Jetten’s Instagram account and concerned government apologies issued this summer for the treatment of Moluccans following their transfer from Indonesia in 1951. Moluccan soldiers had fought with the Dutch army during the colonial war of independence and expected to establish a Republic of the Moluccas. Instead, they were prevented from returning to their homeland, regarded as traitors, and housed in poor conditions without an integration policy.
Jetten visited the former Nazi deportation camp at Westerbork, in the north-east of the Netherlands, where Moluccans were forcibly interned after the Second World War, and published a reflection. One passage read: “Different histories, different forms of injustice. Not to compare them with one another, but to continue telling them alongside one another.” According to the newspaper’s investigation, the passage was generated by a chatbot but published under Jetten’s name.
Pieter Walraven, an expert at the consultancy Public Matters specialising in political influence and strategic communication, told NOS that AI “is a welcome support for organisations and politicians”. But he warned that the situation becomes more complicated “if politicians start forming their opinions based on AI”, describing that development as concerning.
De Volkskrant used the AI-detection programme Pangram to examine Jetten’s communications and, as an additional control, assessed about 700 posts published by the prime minister and D66 before ChatGPT was launched in 2022. None was identified as AI-generated by the same programme.
The investigation extended beyond Jetten. Television programme Nieuwsuur analysed 77 speeches by Defence Minister Dilan Yesilgöx and her State Secretary, Derk Boswijk, identifying 15 as suspicious. Ten contained AI-generated passages, while five consisted almost entirely of AI-generated text. Thirteen of the 15 came from the minister and state secretary, including five speeches that appeared to have been written entirely by AI.
Nieuwsuur also found that at least 12% of the text in speeches by Education Minister Rianne Letschert and her State Secretary, Judith Tielen, had been generated by AI.
So far, Geert Wilders, the Dutch far-right leader, has been the only politician to react to the findings. According to NRC, he wrote on X: “Even the prime minister does not exist; Robbie is fake. #getout”, before deleting the post shortly afterwards.
