“As an experienced psychologist and avid defender of psychology behind human behavior ….” writes the AI author of the Flemish magazine Psychologies. The article about ‘typical human behaviors’ was not written at all by an ‘experienced psychologist’. Is it time for a general guideline for the use of AI (artificial intelligence) within journalism?
Last week unveiled The VRT that three magazines from Belgium used AI used to write articles. More than half of the articles that appeared in 2025 on the website of the Belgian version of the magazine Elle Was written by invented authors with AI-generated profile photos. That the articles were actually generated by a computer was not mentioned. Also Marie Claire,, ” Forbes and Psychologies went wrong.
A manager of Ventures Media, the parent company of the magazines, stated to VRT that it was a test, in which a ‘limited’ number of articles were generated by AI and not edited. As a result of the investigation, the fake profiles have been removed and disclaimers are given in the articles.
With the use of AI in journalism, things went wrong on an international level. Thus the Chicago Sun-Times Earlier this year an artificially generated book list, where non-existent books were found be in the event of it. And the Dutch fashion magazine Harper’s Bazaar was caught on AI use due to incorrect summaries of books in an article.
AI is high on the agenda everywhere due to rapid development. There are also opportunities and useful applications for journalism. Both The Washington Post when Associated Press have been using self-developed AI models for more than 10 years. These models were used to generate short articles and to help the journalists speed up their editorial process.
Nevertheless, Dutch journalism remains cautious about using AI. There is no consensus between Mediahuizen about the scope and the application. Many editors have drawn up an AI policy themselves. This policy is particularly strict with newspapers. In the protocol of de Volkskrant State that the newspaper AI sees as a tool, but that it can never replace the work of a journalist. Also in the policy of NRC and FD His guidelines laid down on the use of AI, where the promise is that AI is never used without human involvement and supervision.
The Journalism Council currently does not use a general regulation for the use of AI. They leave NRC Know the current guidelines to be sufficient to include AI-related issues. This directive does exist in Belgium, and it is forbidden to publish AI-generated articles without mentioning it. The Flemish Association of Journalists (VVJ) therefore says that with ‘growing indignation’ to have taken note of the use of AI in the Flemish magazines.
To trust
The Dutch Association of Journalists (NVJ) also rejects the editorial choices of the responsible Flemish magazines. “The titles that do this harm journalism,” says secretary Thomas Bruning. ,, In the end, journalism is based on trust. The moment you release that principle, you create damage to yourself, but also trust in the profession. ”
Teacher-researcher Nele Goutier, specialized in AI in journalism says that the most important thing is not to lose sight of your audience. “I think that ELLE Belgium Has been carried away in the many possibilities of technology, and has not thought of what the public and the profession expects from them, “she adds.” Does the public want more content? They are already flooded by the amount of online information nowadays. And research shows that people think it is always important that a journalist is always involved. “
“Ai can bring journalism a lot of beautiful things,” says Goutier. “Certainly if it is used as an aid in the editorial process of the journalist or to enrich the experience of the public. Research shows that the audience finds two things very important for AI-Implications: there must always be human control and the journalist must be transparent about the AI use.”
The NVJ also sees this transparency and human control as the most important guideline when using AI in journalism. The association has long been busy informing journalists about the risks, but also the possibilities of AI. “Ultimately, the responsibility lies with the editors themselves. But it is important for all journalists to make clear agreements about this.”
Give various Dutch magazines to NRC To publish AI articles, but AI is increasingly finding its way on the editors. Louise van Nispen, managing director of Linda (which also includes Vogue.nl), says that their editors AI are seen as a powerful means of increasing their ‘creative impact’. There is an AI policy to handle the systems responsibly. Our brands Linda,, ” Linda girls and Vogue.nl So definitely use AI, but not for the development and writing of content. If there is AI content in the magazines, this will be mentioned. ”
The NVJ cannot say whether a general guideline is the solution. “It is difficult to have the credibility for a journalist checked by third parties,” says Bruning. “We can’t do that anyway and it is also not the task of the Journalism Council. A guideline would therefore only be prescribing and not immediately solve a problem.”
Despite the fact that more and more editors are testing with the possibilities of AI, a future in which the AI takes over the journalist is unlikely. “The language models on which AI is based are actually just statistics. They don’t understand the content of what they write,” says Goutier. “If you get people out of journalism, journalism loses its basis of reliability. AI models are currently not enough to trust in content to use journalistic purposes without human control. The new models make even more mistakes than the first models. The human journalist is really needed to provide reliable news.”