Dutch PM Jetten used AI to write social media posts — and defence ministers may have used it for speeches

Dutch Prime Minister Rob Jetten has admitted to using artificial intelligence to write hundreds of his social media posts, a practice revealed by De Volkskrant on 10 September 2026. The same day, a Nieuwsuur analysis of 77 ministerial speeches found that Defence Minister Dilan Yesilgöz and State Secretary Derk Boswijk also appear to have used AI extensively — though the ministry insists it was only an aid.

What happened?

De Volkskrant reported that Jetten, who leads the D66 party, had been using AI to generate posts on platforms like Facebook and Instagram for months. The paper used the software Pangram to analyse the language, which flagged hundreds of posts as likely AI-written. Jetten confirmed the practice to NOS, calling AI “a helper” and saying that “just because you use AI as a tool doesn’t mean a human eye doesn’t look at it.”

One example cited was Jetten’s post about the government’s apology to the Moluccan community earlier this year. The post used language that Pangram identified as machine-generated.

Meanwhile, Nieuwsuur applied Pangram to 77 speeches published on the Rijksoverheid website. Of those, 15 were flagged as suspicious: 10 contained AI-generated fragments, and five appeared to be almost entirely AI-written. Thirteen of those 15 came from Defence Minister Dilan Yesilgöz and State Secretary Derk Boswijk, including all five fully AI-generated speeches.

Speeches from Education Minister Rianne Letschert and State Secretary Judith Tielen also showed partial AI use, but at much lower levels (12% each).

Why does this matter?

For the average person in the Netherlands — especially expats and internationals who rely on clear, trustworthy information from the government — this raises a fundamental question: who is actually speaking?

“It’s essentially automating what press officers and spin doctors used to do,” said Mariken van der Velden, professor of political communication at Vrije Universiteit Amsterdam, in an interview with NOS. “The problem is that AI is often mentioned in the same breath as misinformation. You want to be able to trust information from the prime minister.”

Tom Dobber, assistant professor of political communication at the University of Amsterdam, told Nieuwsuur that the language in the flagged speeches was “uncomfortable and hollow, a bit meaningless — and really typical AI formulations.”

Claes de Vreese, professor of AI and society, added that “language models create a kind of desirable emotion, not necessarily what you actually feel. People will start to wonder: who is the real Jetten? This way, the authenticity of the prime minister erodes.”

The transparency gap

D66 — Jetten’s party — supported the European Union’s AI Act, which emphasises the importance of transparent AI use. Yet the party did not disclose that it was using AI for Jetten’s posts. This inconsistency has not gone unnoticed.

Van der Velden argues that the lack of disclosure is the real issue: “If you use AI, you should say so. Otherwise, you undermine the credibility of the office.”

What does this mean for you?

If you live in the Netherlands on a residence permit, you may rely on official government communications — from ministerial speeches to social media updates — to understand policy changes, deadlines, or your rights. If those communications are written by AI without disclosure, it becomes harder to assess their reliability.

The broader concern is that AI-generated language can sound convincing but may lack nuance, context, or the personal accountability that comes with a human author. For example, a speech about defence policy or an apology to a community could be crafted to sound appropriate without reflecting genuine intent.

What happens next?

As of now, Jetten has not committed to mandatory disclosure when AI is used in official communications. The Defence Ministry denies that AI is used to write complete speeches, stating it is only a help tool. But the evidence from Pangram — and the experts who reviewed it — suggests otherwise.

This story is still developing. Questions remain about how widespread AI use is across other government departments, and whether the lack of transparency violates the spirit of the EU AI Act that D66 supported.

For now, the advice for anyone following Dutch politics is simple: read official statements with a critical eye, and pay attention to whether disclosure labels start appearing.

---

Sources: NOS – Premier Jetten liet zijn berichten door AI schrijven, NOS – Jetten deed het op sociale media, deze kabinetsleden schreven hele speeches met AI, NOS – Wekdienst 11/9