This week, misleading AI-generated videos and images related to the situation in Venezuela and its ousted leader Nicolas Maduro garnered millions of views across major platforms. Simultaneously, mainstream social media platforms and their AI detection tools are struggling to keep pace with the rapid advancement of AI technology, a problem that becomes particularly acute during breaking news events.
Following the US military operation to overthrow Venezuelan leader Nicolas Maduro, a series of AI-generated videos purporting to show Venezuelan citizens celebrating in the streets quickly went viral on social media. These AI-synthesized clips depicted scenes of jubilant crowds and accumulated millions of views on mainstream platforms such as TikTok, Instagram, and X. One of the earliest and most widely circulated video clips on X was posted by an account named "Wall Street Apes," which boasts over one million followers on the platform. The post showed several Venezuelan citizens weeping with joy, with individuals in the footage expressing gratitude to the United States and President Donald Trump for overthrowing Maduro. The video was subsequently flagged by X's "Community Notes" feature, a user-powered, crowdsourced fact-checking mechanism that allows users to add contextual information to posts they deem misleading. The note stated: "This video is AI-generated and is currently being spread as a real event, aiming to mislead the public." This video has been viewed over 5.6 million times and shared by at least 38,000 accounts, including business magnate Elon Musk, who later deleted his repost. CNBC could not independently verify the origin of the video, but fact-checkers from the BBC and AFP stated that the earliest known version of the video appeared on a TikTok account "@USCuriousThinkTank," which frequently posts AI-generated content. In fact, even before such videos emerged, AI-generated images depicting Maduro being detained by US forces were circulating online before the Trump administration released authentic photos of Maduro's arrest. The former Venezuelan president was arrested on January 3, 2026, following US airstrikes and a ground raid. This military operation dominated global headlines at the start of the new year. Beyond these AI-generated videos, AFP's fact-checking team has flagged numerous other misleading pieces of content related to Maduro's downfall, including video clips that misrepresented footage of celebrating crowds in Chile as scenes from Venezuela. The proliferation of misinformation during major news events is not a new phenomenon. Similar false or misleading content has emerged previously during the Israel-Hamas conflict and the Russia-Ukraine war. However, the AI-generated content related to the recent events in Venezuela, with its vast reach and high degree of realism, serves as a stark example of AI technology increasingly being weaponized as a tool for disinformation. The emergence of AI content generation platforms like Sora and Midjourney has made it significantly easier to rapidly produce hyper-realistic videos that can be passed off as genuine amidst the information chaos of breaking news. The creators of such content often aim to amplify specific political narratives or sow confusion among a global audience. Last year, an AI-generated video also went viral online, featuring several women lamenting the loss of their Supplemental Nutrition Assistance Program benefits during a government shutdown. The video even fooled Fox News, which initially published an article treating it as a real event before later removing the report. Given the intensification of such trends, social media companies are facing growing pressure to enhance their efforts in clearly labeling potentially misleading AI-generated content. Last year, the Indian government proposed legislation mandating the labeling of AI-generated content, while Spain introduced regulations that could impose fines of up to 35 million euros for unlabeled AI content. In response to mounting concerns, major social media platforms, including TikTok and Meta, have rolled out tools for detecting and labeling AI content, but with mixed results. CNBC found that some videos being shared as celebrations in Venezuela on TikTok had already been labeled as AI-generated. Conversely, X primarily relies on its "Community Notes" feature for content labeling. Critics argue that this mechanism is often too slow to react, failing to prevent the widespread dissemination of AI disinformation before it is identified. Adam Mosseri, who oversees Instagram and Threads, acknowledged this challenge facing the social media industry in a recent post. He stated, "All major platforms are making efforts to identify AI-generated content, but as AI gets better at mimicking reality, our ability to identify it will likely diminish." He added, "Increasingly, many people, including myself, believe that watermarking authentic content digitally will be a more viable solution than labeling fake content."
Comments