Thousands of Dutch citizens scrambling to leave the Middle East were drawn into a high-stakes narrative of evacuation flights that never took off, after the Netherlands’ largest newspaper, De Telegraaf, ran an interview with a woman claiming to organise her own chartered trips from Dubai. The report, published online on March 5 under the headline “Dutch people in the Middle East feel abandoned by the government: We just rented a plane ourselves,” detailed the efforts of a group renting an Airbus A321 to transport fellow Dutch citizens home at a cost of €1,600 (US$ 1,850) per ticket.
The story emerged in the tense aftermath of Iranian missile and drone strikes across the Gulf, following US-Israeli military actions, when thousands of Dutch citizens were seeking urgent ways to leave the region. During a television interview, the Dutch minister of foreign affairs faced questions directly inspired by the article, assuring the public that official government repatriation efforts were underway.
The article featured interviews with several Dutch travelers, including Tamara Harema, who claimed to have faced repeated flight cancellations and frustrations with official channels. Harema described how she and a group of fellow citizens were organising buses and chartering an Airbus A321 for private evacuation. “The first plane is already full, so we’re organising a second flight. Stranded travellers can contact us,” she was quoted as saying.
However, investigative work by Bellingcat raises serious doubts about the story’s authenticity. The original photo accompanying Harema’s interview appears to exhibit numerous artefacts associated with generative AI. The image, showing a view of Dubai’s skyline with the Burj Khalifa in the background, contains inconsistencies when compared with both historical imagery and the architectural realities of the tower. Mechanical floors are misaligned, key water features at the base of the building are missing, and the perspective of the hotel room does not match any known angles from that side of the tower. Additionally, irregularities in furniture and details of Harema’s clothing and jewelry suggest AI manipulation.
Flight data further undermines the narrative. Harema claimed that buses would collect passengers from Dubai on March 7 to board a 232-seat Airbus A321 departing from Muscat, Oman, to the Netherlands. Analysis of flight tracking information revealed that no such flights took place. While multiple A321s departed Muscat on March 7 and 8, none were headed to the Netherlands, and only government-organised repatriation or scheduled Oman Air services arrived in Amsterdam. Two A321s recorded on the ground at Muscat were later tracked flying to Rome and operating regional routes, with no evidence of a Muscat-to-Amsterdam evacuation flight.
Following Bellingcat’s inquiries, De Telegraaf removed Harema’s photo from the article, adding a note stating that the image “likely did not meet our journalistic guidelines.” Joost de Haas, deputy editor-in-chief, clarified that Harema had been contacted through Chizki Loonstein, a lawyer living in Dubai, who provided her contact details. De Haas explained that quotes were extracted from communications with Harema after Loonstein introduced her to reporters, but further verification was limited.
Loonstein, whose background includes a 2025 report noting the collapse of his legal firm in the Netherlands and subsequent emigration to Dubai, confirmed knowing Harema but refused to provide further comment beyond the initial introduction. Bellingcat verified the inconsistencies in both the photograph and flight information through open-source research, emphasizing the unusual combination of AI imagery and nonexistent flight claims.
The incident highlights the growing risks of AI-generated media contributing to panic during crises. Dutch citizens, already anxious due to regional instability, were presented with a story that appeared credible but could not be substantiated. The removal of the photo and the lack of verifiable flights indicate that readers were, at least partially, misled by digital manipulation. Bellingcat’s work in uncovering these inconsistencies underscores the role of investigative open-source reporting in verifying high-stakes information.
Experts note that AI-generated content is increasingly difficult to detect without careful cross-referencing of visual and factual details. In Harema’s image, for example, even subtle anomalies in architectural features, clothing, and accessories provided clues to the artificial nature of the photo. Similarly, discrepancies between claimed flight schedules and publicly available aviation data revealed the nonexistence of the alleged evacuation services.
The De Telegraaf incident also illustrates challenges for traditional news outlets operating under intense deadlines. The temptation to publish sensational reports, especially amid high anxiety and rapidly evolving events, can lead to the amplification of unverified claims. In this case, the story’s timing—amid widespread concerns over regional security—lent additional urgency and perceived credibility to the narrative, despite its flaws.
As governments and citizens navigate the growing prevalence of AI-generated misinformation, this episode serves as a cautionary tale. Bellingcat’s investigation demonstrates the importance of rigorous verification and the application of open-source tools in separating fact from fiction. While thousands of Dutch travelers continued to rely on official repatriation channels, the false narrative of privately organised evacuation flights highlighted the real-world consequences of disinformation amplified through traditional media channels.
