A blonde films herself at a racetrack and says: "If you support Trump, you just made a friend." A brunette films herself at a stadium and says the exact same thing. A redhead is at a basketball court. Same line. Same grammatically awkward caption: "I'm new here and love God, America, and Trump!!"

All three women are artificial intelligence.

The New York Times reported today that hundreds of AI-generated pro-Trump influencer accounts have surged across TikTok, Instagram, Facebook, and YouTube in the months leading up to the 2026 midterm elections. The accounts feature avatars that post at a rapid pace about the "radical left" and "America First." They tend to appear as ordinary, very attractive men and women, gazing into the camera while talking about the war in Iran, abortion, or Bad Bunny.

President Trump reposted content from at least one of these accounts: a platinum blond avatar making unfounded claims about California's governor.

And that is not the only front. The National Republican Senatorial Committee released a deepfake video of Texas state Rep. James Talarico, the Democratic Senate candidate, appearing to "read" his own years-old social media posts. The words "AI Generated" appear in tiny, all-caps text at the bottom right corner of the video. Most viewers will never see them.

The Apocalypse That Showed Up Late

In 2024, everybody braced for an AI deepfake apocalypse in elections. It did not arrive. The fakes were clumsy. Hands had extra fingers. Short-form videos looked stiff. Campaign operatives feared public blowback more than the fakes themselves. The robocall imitating Joe Biden that told New Hampshire primary voters to stay home had limited real impact.

Two years later, the technology caught up and the norms collapsed. The quality is better. The volume is orders of magnitude higher. And the political ecosystem has changed so fundamentally that nobody seems to care anymore. In January, the White House itself posted an altered photograph of an African American protester with darkened skin. Anti-ICE activists used AI to "unmask" immigration officers. Both parties are doing it. Both sides have stopped pretending they are not.

The American Prospect published a deep investigation this week showing the full scale of the problem. Adam Rose, a fellow at the Starling Lab for Data Integrity at Stanford and USC, laid out the real danger: it is not just that fake content exists. It is that fake content makes real content less trustworthy.

"Without visual evidence, we lack the ability to understand what is happening and to make judgments both in a court of public opinion and in a court of law," Rose told the Prospect. "If people in either of those courts do not trust the evidence, it makes it very difficult to function as a civil society."

That is the "liar's dividend": a world where real evidence can be dismissed as AI-generated, and nobody can agree on what is real anymore.

Both Sides of the Atlantic, Same Problem

It is not just America. The UK Electoral Commission just launched a deepfake detection scheme ahead of May local elections, warning of "increasingly sophisticated technology" that could undermine voter trust. Britain is trying to get ahead of the problem. America is not even trying.

The prospects for federal action are dim. The Trump administration has taken a hard line against any kind of AI regulation. The few AI deepfake laws that exist are state-level patchworks. The TAKE IT DOWN Act produced its first conviction earlier this month, but that case involved intimate images, not political content.

Meanwhile, the scale of the operation keeps growing. Every AI fake influencer that gains a following is not just spreading messages. It is training the algorithm to show similar content to similar audiences. The AI is not just creating the propaganda. It is optimizing the distribution.

The 2024 AI apocalypse was a false alarm. The 2026 version is already here, and it looks like an attractive person on TikTok telling you to support their candidate. You will not know it is fake. That is the entire point.

Sources: The New York Times, The American Prospect, The Independent, UK Electoral Commission, Stanford Starling Lab.