Loosely organized groups of self-described right-wing trolls and “memers” have long posted flattering portraits of President Donald Trump, portraying him as a war hero, a rock star or a muscular comic book vigilante. Their pro-Trump content, for the most part, was not heavily viewed.

Trump is bringing them into the mainstream.

Trump has been sharing memes online since his first term, but artificial intelligence has made it far easier to generate them. In recent months, Trump has reposted more than 20 AI-generated images and videos, some of them produced by members of online communities dedicated to sharing their experiments with the technology. One video he shared on Truth Social showed him doffing a crown like a king. Another depicted him as a fighter pilot dropping what looked like excrement on protesters.

“WOWWW!!,” the account behind the fake pilot video posted after Trump amplified it on Truth Social. “This is incredible.”

Communities Devoted to Praising Trump

The New York Times examined the online ecosystem of AI aficionados who have organized into communities devoted to praising the president. Hundreds of users, posting anonymously each day, have produced thousands of artificial intelligence-powered videos and images displaying their fondness for the Trump administration and mocking the president’s enemies. Their work is often crude and sometimes racist.

Trump has, at times, reposted their work to his social feeds, catapulting the memes from the periphery of political discourse to its center and upending traditional pathways to influence.

The White House did not comment on the president’s reposts, although a spokesperson for the administration recently praised Trump’s use of social media to “communicate directly with the American people.”

At least a dozen of these meme groups are active on the social media platform X and Discord, a messaging service popular with gamers. Their creators have met online for years and congregated in different forums, posting memes on fringe internet forums and pushing cryptocurrencies that riff on Trump’s brand. Often, those videos perpetuate harmful stereotypes about immigrant communities; many utilize racial slurs.

Even now, the groups remain obscure. They organize in private channels to share their work and troubleshoot issues. Most of their posts simply seek an audience, though some of the collectives offer their services to political campaigns or brands for a fee.

But artificial intelligence has made their content more visceral and viral. Services that can create realistic AI videos — such as OpenAI’s new tool Sora — have become far more powerful and accessible over the past year. Users can now create almost anything they can imagine by typing instructions into a text box, often for little or no cost.

Meme Collectives Expand Reach

Lately, meme collectives are expanding their reach, whether because of increased public interest in AI content, or a result of landing on the administration’s radar. Together as a unit, the members of Green Frog Labs, one of the loosely-organized groups on X, claim more than 300,000 followers on X and more than 100 million views of their videos, according to the group’s estimates.

“We really have some of the best memers,” said Greg Scott, an entrepreneur who set up the group earlier this year, in an interview with The New York Times.

The group has more than 90 members ranging in age from 20 to 60 (Scott is in his 40s) and quickly found success producing AI videos mocking one Hollywood actress and replacing people from news segments with AI-generated talking babies. Creators crafted videos of Trump as Patrick Bateman, the lead character of the film “American Psycho,” preparing an agenda of “vengeance and retribution,” and also as a Rambo-type figure firing sombreros out of a machine gun.

Recently, Green Frog members produced dozens of videos targeting Zohran Mamdani, the mayor-elect of New York, often leaning on religious and racial stereotypes.

“I understand people would be offended, might find it hateful,” Scott said. “It could come across as that to some people, but we’re just trying to make people laugh.”

Members of Green Frog and other groups have congregated on Discord, largely anonymously, to swap tactics for bypassing content restrictions on popular AI apps, according to messages reviewed by the Times. They have tested the limits of mainstream tools such as Sora, trying to evade guardrails meant to prevent the app from generating depictions of notable figures like Trump. As the groups’ efforts become more sophisticated, some have evolved into video production businesses.

“Little by little we’re figuring out how to use President Trump on the new Sora,” one creator posted on X in late October, alongside a video depicting an AI replica of Trump celebrating the end of funding for the Supplemental Nutrition Assistance Program, or SNAP.

The Trump White House has eagerly embraced AI as a propaganda tool, following the president’s lead in posting artificially generated content showing various Democrats in sombreros — material that Rep. Hakeem Jeffries, the House minority leader and a frequent target of the videos, has called racist and bigoted.

In March, the official White House account on X faced backlash after sharing an AI-generated image of immigration officers arresting a woman who was previously convicted of trafficking the drug fentanyl. Kaelan Dorr, the deputy communications director at the White House, wrote on X that people seemed more upset about the AI image than they were about the fentanyl crisis.

“The arrests will continue,” he wrote. “The memes will continue.”

—

This article originally appeared in The New York Times.

By Stuart A. Thompson and Tiffany Hsu
c. 2025 The New York Times Company