During the recent Dutch elections, fact-checking efforts intensified across the media landscape. Platforms like BOOS, led by Tim Hofman, actively debunked migration myths circulating on Instagram and TikTok. De Volkskrant revived its "Klopt dit wel?” section to scrutinize political claims, while established fact-checking organizations such as Nieuwscheckers (Leiden University) and AFP Fact Check continued their relentless work to verify public information.

When generative AI first appeared, many feared it would unleash a flood of fake content, overwhelming fact-checkers. However, recent data from the European Digital Media Observatory (EDMO) shows that generative AI has been a factor in only 1.3% of fact-checks conducted by its member organizations. Earlier research also found that AI-generated content likely had no significant impact on the 2024 European elections.

Yet, researchers caution that the absence of widespread AI-driven misinformation today does not mean Europe is immune in future elections. The technology’s capacity to mimic credible sources remains a ticking time bomb.

A recent investigation by BNR Nieuwsradio demonstrated how OpenAI’s Sora 2, the company’s latest video-generation model, can easily produce fake news videos that look indistinguishable from authentic broadcasts by Dutch outlets like BNR and NOS. One fabricated clip even appeared to show the VVD celebrating a major election victory—completely fictional, yet alarmingly realistic.

OpenAI has reportedly built safeguards into Sora 2 to prevent impersonation of English-language media, but Dutch outlets seem to have been overlooked.

The incident underscores an uncomfortable truth: while AI has not yet overwhelmed fact-checkers, it has the power to do so at any moment. Vigilance, transparency, and continued investment in verification tools are now essential pillars of responsible journalism in the AI era.