Foto eines älteren französischen Ehepaars, das während der Unruhen "entkleidet und ausgeraubt" wurde, ist von AI

Top Democratic leaders on the United States congressional intelligence committees disclosed on Wednesday that they received a classified briefing indicating Russia is actively waging a misinformation campaign against Democratic candidates ahead of next month’s midterm elections. Senator Mark Warner of Virginia and Representative Jim Himes of Connecticut stated that the intelligence briefing, requested from Director of National Intelligence Jay Clayton, revealed Moscow is employing tactics similar to those observed in American elections in 2016, 2020, and 2024. Warner characterized the current situation as Russia targeting Democrats specifically, noting that the campaign mirrors previous efforts to undermine electoral integrity. The disclosures came as President Donald Trump has recently eased economic sanctions on Russia and increased diplomatic pressure on Ukraine to cease hostilities with Russian forces. Warner and Himes emphasized that they were restricted from sharing specific details of the briefing due to its classified nature. However, they identified specific targets of the disinformation effort. Warner cited news reports alleging that Texas Senate candidate James Talarico and Georgia Senator Jon Ossoff, who is seeking re-election, have been subjected to fake advertisements generated using artificial intelligence. Himes described the Russian approach as chaotic, stating that Moscow is attempting to sow distrust in voting machines and general confusion. He summarized the strategy as consistent with past behavior, describing it as the same recurring pattern of interference. Neither the White House nor the Office of the Director of National Intelligence responded immediately to requests for comment regarding the briefing. The Russian embassy also did not immediately respond to inquiries, though Moscow has historically denied allegations of meddling in U.S. elections. While the primary focus of the legislative disclosure concerns domestic political security, the broader landscape of misinformation currently involves significant international cross-currents. In France, a separate wave of viral imagery related to ongoing civil unrest has required distinct verification processes. Mass protests have swept across France since late September, initially driven by students protesting poor classroom and learning conditions. These demonstrations escalated on October 17, when a national day of action supported by student groups, labor unions, and the Yellow Vests movement resulted in several violent incidents. Historical structures, including the Limoges Town Hall, suffered extensive fire damage during these events. Authentic footage of the destruction exists, but it has become entangled with fabricated visuals circulating on social media platforms. One prominent example of this fabrication is an image depicting protesters waving Palestinian flags and throwing fireworks at the base of the Eiffel Tower. This image has been shared online with captions claiming that “France is approaching civil war” and that “Paris has fallen.” Verification teams determined that this image is not real. It contains an invisible SynthID watermark, which indicates it was created or edited using OpenAI tools, including ChatGPT, the OpenAI API, or Codex. Visual analysis further revealed inconsistencies, such as figures in the background blending into one another and into the surrounding smoke. Tracing the image back to its origin, researchers linked it to a satirical article published on October 6. That article humorously posited that Palestinians had been granted their own country within the borders of France, providing the context for the fictional imagery now being misinterpreted as evidence of actual conflict. Another widespread false narrative involves a photograph of an elderly couple appearing to be stripped and robbed during the riots. This image has accumulated millions of views, particularly on the platform X, where it was amplified significantly by far-right accounts. Captions accompanying the photo accused “far-left and migrant rioters” of stripping the couple in the street. Some users described the scene as the most horrific incident of October 17, while others used it to claim that France had fallen and warned that the rest of Europe and the United Kingdom would face similar fates. Additional inflammatory claims suggested the couple was Jewish, with some users accusing President Emmanuel Macron of failing the Israeli people and comparing the alleged attack to a 1941 pogrom in Ukraine. American far-right commentator Alex Jones shared one variation of the image, attributing the violence to a “criminal consortium of Muslims and communists.” Despite the virality of these claims, no official reports confirm that such an attack occurred. The public prosecutor's office informed the French newspaper Libération on October 19 that it had received no reports of an incident matching the description. Technical analysis confirmed the artificial origin of the photographs. When run through Google’s SynthID detector, both variations of the image were flagged as having been made or edited using Google’s AI tools. Visual anomalies support this finding; in one version, the elderly woman appears to have three fingers and a large, deformed thumb, while a masked individual nearby appears to have two heads attached to a single body. A bollard depicted in the image does not match types typically found in Paris. Researchers traced the earliest version of the image to a post on X dated October 17, which was part of a montage flagged by the platform as AI-generated. The montage was published by a pro-Israel account named Infos-Israel.news. The images also circulated beyond social media, appearing on the Instagram page of French digital media outlet Sirenes and on the Israeli right-wing news channel C14, which published one under the headline “Islam Conquers Europe.” The persistent white border present in all shared versions suggests the original montage as the common source.