AI-generated content falsely links stars to anti-Somali messages
- Published on January 21, 2026 at 21:30
- 2 min read
- By Sahas WIJEWARDENE, AFP USA
Inflammatory posts circulating on Facebook claimed popular music stars had attacked Somali immigrants and called out Minnesota Representative Ilhan Omar. But the claims are false and part of a wider trend of AI-generated clickbait posts that recycle similar content under the names of different celebrities, often spread by accounts operating outside the United States.
"Our country would be safer without Somali immigrants -- Starting with Ilhan Omar!" exclaims a January 7, 2026 Facebook post, coupled with pictures of country music star Brandon Lake alongside Omar.
Near identical posts give more excerpts from Lake's alleged speech which suggests people should start speaking up for the "silent majority."
"This country welcomes people in good faith," Lake supposedly said. "But what we get in return -- from some -- is contempt for our culture, our values, and our Constitution."
The posts encouraged users to read more in an article linked in the comment section.
AFP found duplicates of the text swapping out Lake for other country musicians such as Jason Aldean, Reba McEntire and Gretchen Wilson, along with singer Barbra Streisand, and rapper 50 Cent.
The Trump administration has latched onto news of large-scale public benefit fraud to target Omar and the large Somali community in Minnesota (archived here). But misinformation about the Minnesota Democrat and her constituents also regularly circulates.
On January 13, the United States said it would end a special protected status for Somalis, telling them they must leave the country by mid-March under an escalating crackdown on the community.
The president has called Omar "garbage," but keyword searches on Google yielded no credible evidence that any of the stars mentioned in the posts made the statement about the first African refugee elected to Congress.
Many of the Facebook accounts spreading the posts claimed to be based at US addresses but were in fact operated from Vietnam, a common location for spam networks seeking to profit off of engagement.
The account which posted on Lake is called Tune Nest, and has more than 20,000 followers.
Further review revealed that the articles cited in the posts included unnatural English phrasing and homoglyphs -- characters that look similar but are encoded differently -- pointing to the use of artificial intelligence.
"The reason why these websites use homoglyphs on their text is to avoid being detected as having been AI-written and also to bypass automatic fake news detectors," said Aldan Creo, a large language model researcher at the University of California, San Diego (archived here), in a previous interview with AFP on December 26, 2025.
Representatives of public figures targeted in the posts, such as Lake, Wilson and Alden did not respond to AFP's requests for comment.
AFP previously debunked similar AI slop here, here, and here.
Copyright © AFP 2017-2026. Any commercial use of this content requires a subscription. Click here to find out more.
Is there content that you would like AFP to fact-check? Get in touch.
Contact us
