AI Chatbots Flag "Uncivilized Media" as Antisemitic, Yet Platforms Allow It to Spread
Four leading AI systems identified the Qatari outlet's "global Israeli colonizers" video as antisemitic, raising questions about whether social media platforms failed to detect it, or chose not to act
Jewish Onliner is an independent publication. If you find our work valuable, please consider becoming a paid subscriber. You can also subscribe to us on WhatsApp and Telegram. Follow us on YouTube, TikTok, Instagram, Discord, and X.
Uncivilized Media is a Qatari outlet that presents itself as an American youth-culture media brand. Its newest video argues that Israeli travelers carry violence and colonization “in the DNA” — rhetoric that every major AI chatbot recognizes as antisemitic within seconds of being asked. The video has racked up hundreds of thousands of views anyway, cross-posted across multiple platforms, without any of them appearing to intervene.
A Jewish Onliner review of Uncivilized Media found that when shown the transcript of the outlet’s viral video portraying Israeli travelers as “global colonizers,” four major AI chatbots identified it as antisemitic without needing any specialized prompting. YouTube, where the video is hosted, and the social accounts where Uncivilized cross-posts its clips, apply no comparable standard. The video remains up and continues to accumulate views.
Every Model Flagged the Content
When provided with the transcript of Uncivilized Media’s “How Israelis Are Colonizing the World” video, all four models identified antisemitic or hateful elements in its rhetoric:
ChatGPT concluded that it “crosses into antisemitism by portraying standard travel and emigration as a sinister global plot” and claiming violence is “in the DNA” of Israelis.
Claude said the video uses “essentialist language about a people” by describing violence as living “in the DNA” of Israeli settlers and framing voluntary migration to Italy as “colonizing the world,” language it said many would flag as antisemitic.
Grok determined that it “crosses into antisemitic and hateful territory by portraying Israelis as a single, inherently colonial group.”
Perplexity said it “uses rhetoric that reads as antisemitic and hateful,” including “sweeping claims about Israelis/Jews as a group.”
None of this required specialized context or training most viewers would lack. It took a transcript and a direct question. That is the baseline the platforms hosting and distributing the video should be measured against.
Qatari Propaganda, Not Grassroots Media
Uncivilized Media, as Jewish Onliner previously reported, is registered in Qatar’s state-backed Media City hub and led by Salem Barahmeh, a veteran of PLO diplomacy who previously worked for Portland Communications, a firm contracted by Qatar for public relations. The outlet’s co-founders have backgrounds at Al Jazeera and the Doha Film Institute.
The video itself argues that Israeli tourists in Thailand, families relocating to Italy, and backpackers in India are all expressions of the same “colonizing” project. On YouTube, it passed 500,000 views within six days of posting, drawing more than 42,000 likes and 10,000 comments. Uncivilized also runs an active TikTok account with 74,000 followers, and an Instagram account with 203,000 followers — where it regularly cross-posts clips from its YouTube catalog to reach separate audiences on each platform.
On Instagram, Uncivilized’s “global colonizers” post has garnered almost 18,000 likes in just one week.
YouTube’s AI Is Built to Catch This
YouTube says hate speech enforcement does not rely on user reports alone. The company has stated that it relies on “a combination of people and machine learning technologies” to enforce its rules, backed by a review team of more than 20,000 people. Its Hate Speech policy bars content that promotes hatred against groups on the basis of nationality, ethnicity, or religion. “How Israelis Are Colonizing the World” remained online and gaining views as of this writing.
AI chatbots show a parallel blind spot. Asked why ChatGPT continues to reference Uncivilized Media neutrally despite flagging its content as antisemitic, an OpenAI spokesperson pointed to the company’s Model Spec, which permits ChatGPT to analyze harmful material without endorsing it but does not address how the system decides which outlets to treat as credible sources.
Anthropic, Google, X, and Perplexity did not respond to requests for comment.
Other Social Media Platforms Make the Same Claims
Meta makes similar claims for Facebook and Instagram. The company’s Transparency Center says its enforcement AI now covers languages spoken by 98 percent of people online, up from about 80 languages previously, which it says lets it “apply policies more accurately and consistently across billions of pieces of content.”
TikTok claims the strongest detection performance of the social media platforms. The company says 96.3 percent of the content it removed for hate speech violations in the fourth quarter of 2025 was caught before any user reported it. Uncivilized maintains an active TikTok account that regularly cross-posts commentary from its YouTube catalog to a separate audience there.
Differing Enforcement Philosophies, Same Result
Across the differing enforcement philosophies, the same result remains: content that took independent AI systems seconds to identify as antisemitic kept spreading.
Uncivilized Media’s video was never a hard case. Four independent AI systems identified it as antisemitic the moment they were asked, using nothing more than a transcript. The platforms carrying it — YouTube, Meta, TikTok, and X — have spent years building, and publicizing, AI systems built for exactly that kind of judgment. They let the video reach hundreds of thousands of people anyway.