Urgent.News

What's breaking now, across thousands of outlets.

AI

Nearly half of Dutch MPs use AI to write speeches in parliament

Nearly half of Dutch MPs used artificial intelligence to help write speeches and statements delivered in parliament, according to an...

Nearly half of Dutch MPs use AI to write speeches in parliament

Nearly half of the Dutch Members of Parliament (MPs) utilized artificial intelligence (AI) to assist in writing speeches and statements delivered in parliament, according to an investigation using the Pangram AI detection tool. Out of the 150 MPs from various political parties, 68 individuals employed AI at least once, primarily in committee hearings.

De Groene Amsterdammer and Investico conducted this analysis, revealing that most MPs admitted to using AI tools after being questioned by the publication. Out of these, former Pro MP Esmah Lahlah, now a member of Amsterdam city council’s executive, was found to use AI in 4 debates on poverty and the jobseekers' law. Lahlah explained that she resorted to AI to condense conversations from her working visits into speaking time.

While the majority of MPs crafted their speeches, more than half of the contributions to committee hearings exhibited AI usage. Notable right-wing party users included Rachel van Meetelen of the PVV and Gidi Markuszower’s DNA group of former PVV MPs. Socialists (SP), animal rights party PvdD, and Volt reported no AI usage. Centre-right CDA member Harmen Krul admitted using AI for editing speeches, not generating content.

Prime Minister Rob Jetten also admitted using AI for social media posts but defended it by stating he's not alone in employing AI for betterment. Some MPs dismissed the concerns raised by the latest analysis, comparing AI usage to commonplace tools like calculators and mobile phones.

Written by urgent.news from DutchNews's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dutchnews.nl →

More in AI

Your Java LLM agent works in the demo. Friday night in production is another story.

Friday, 6:40 PM. Picture an agent that handles customer refunds. In the demo it was perfect: it reads the request, checks the order, calls the refund tool and returns clean JSON. Everyone is happy.

  • LLM agent caused 503 errors on Friday night, leading to request timeouts
  • Model wrapped JSON responses in Markdown, breaking downstream parsing
  • ARA (Agent Runtime Architecture) provides rules, controls for production

More from Wednesday 21 October →