On November 9, 2026, Palabra with the National Association of Hispanic Journalists (NAHJ) reported that AI-generated videos depicting fake encounters with U.S. Immigration and Customs Enforcement (ICE) had drawn millions of views on Facebook. The investigation also described AI-made images presented in the style of Latino historical photographs. A separate fact-check offers practical ways to examine a suspicious clip, but no single visual clue settles whether a video is authentic.
What the reported ICE clips depicted
Palabra reported more than 8.5 million views for a Facebook Reel showing a female New York City police officer confronting an ICE agent. A separate clip, depicting a Mexican woman striking an agent, had nearly 6 million views, according to the investigation. Those figures describe two different videos and were reported on November 9; they are not current view counts.
Palabra also reported that an account named Mike Wayne posted thousands of AI-generated fake ICE encounters on Facebook and Instagram, gathering tens of millions of views across its videos. The investigation named the account but did not identify who runs it.
AI-made imagery styled as Latino history
The investigation described three Facebook and Instagram pages—Lost Mexican America, Raices Mexicanas MX, and Mexican American Roots—that published AI-generated images resembling historical photographs. Some posts lacked AI labels or disclaimers, and the images could contain errors, omissions, or half-truths, Palabra reported.
Palabra said the three pages had hundreds of thousands of followers combined and published dozens of posts a day. It reported more than 90,000 followers for Mexican American Roots on each of its Facebook and Instagram pages. These figures were reported by the investigation, not presented as an audited audience count.
How Facebook and Instagram can extend a post’s reach
Palabra described Meta’s recommendation systems as boosting posts that receive early engagement, including by showing them to people who do not follow the account. The investigation also described sharing among relatives, friends, and community groups as another way misleading posts can circulate.
Those distribution routes matter because a video can arrive in a feed without a viewer knowing who made it or where it first appeared. Meta did not respond to Palabra’s requests for comment about its handling of AI-generated misinformation.
How to examine a suspected AI video
A fact-check published by Inquirer carrying DW Fact Check recommends looking for several clues together: distorted or garbled text and logos, unnatural movement, dialogue that does not fit the scene, implausible location details, and identifiers associated with AI tools. It also recommends checking the account that posted the clip, looking for coverage from reputable news organizations, and searching for additional footage or another angle.
These checks are prompts for further investigation, not proof on their own. Generative tools can advance faster than detection systems, and a visual oddity or an automated detector’s result cannot establish a clip’s origin by itself.
When a video feels wrong, check who posted it and look for independent coverage or another angle before sharing it.