Listen to the article
Artificial intelligence researchers are sounding the alarm about a concerning trend: online content creators are not only generating fake images and videos of well-known public figures but are now fabricating entirely fictional people and placing them in military contexts for financial gain and propaganda purposes.
Among these fabricated personas are sexualized AI-generated women in military uniforms who have attracted massive followings online. These digital avatars help reinforce idealized images of political figures, including Donald Trump, even when viewers recognize the content isn’t authentic.
“We are blending the lines between political cartoons and reality,” explains Daniel Schiff, assistant professor of technology policy at Purdue University and co-director of the Governance and Responsible AI Lab (Grail). “A lot of people feel like these images or videos or the stories they convey, feel true.”
The proliferation of political deepfakes has accelerated dramatically in recent years. Since early 2025, Grail has documented more than 1,000 English-language social media posts featuring fake images or videos of prominent political figures and politically significant events. This equals the total number recorded during the previous eight years combined.
This surge stems largely from improvements in generative AI technology, making it “trivially easy to generate a scene that looks pretty realistic and to place real individuals into scenes,” according to Sam Gregory, executive director of Witness, an organization focused on human rights and combating deceptive AI.
The creation of completely fictional avatars represents an even more complex challenge. One notable example emerged in December 2025 when an Instagram account for “Jessica Foster” – an AI-generated blonde woman typically portrayed in U.S. military uniform – began posting images of herself in military settings. Some photos showed her walking beside Trump on an airport tarmac while wearing high heels.
The Foster account, which accumulated more than one million Instagram followers, strategically featured prominent foot imagery and was linked to an OnlyFans account where visitors could purchase supposed foot photos of Foster. The account has since been removed.
“A lot of the AI-generation is to basically get clicks and money or to drive people to a more lucrative place,” Gregory noted.
These tools serve political purposes as well. During conflict in Iran, social media was flooded with fake videos of female Iranian soldiers saying, “Habibi, come to Iran.” The fabrication was evident as Iran prohibits women from serving in combat roles.
Another example includes an AI-generated female police officer with over 26,000 TikTok followers who appears in videos supporting Trump’s deportation policies. During the 2024 election, Trump shared AI-generated images depicting Taylor Swift fans as his supporters. According to Grail’s database, Trump and the White House have shared at least 18 deepfakes on social media since 2024.
The phenomenon crosses political lines. California Governor Gavin Newsom, widely expected to run for president in 2028, has begun sharing deepfakes targeting Trump, including one showing the president smiling at a hologram of Jeffrey Epstein.
Researchers emphasize that these political deepfakes remain persuasive even when consumers recognize they aren’t authentic. Gregory explained that the Foster persona’s implausible scenarios – incorrect military badges, inexplicable interactions with world leaders – don’t diminish their appeal because “people aren’t necessarily looking for things that are real; they are looking for things that represent their beliefs.”
Valerie Wirtschafter, a Brookings Institution fellow specializing in AI and emerging technology, warns that deepfakes make people less likely to reconsider their existing beliefs. These fabrications serve as “just another layer added on in terms of this process of reinforcing, rather than revisiting, what people believe is true.”
The situation may deteriorate further. Technology used to create these fictional personas could produce what researchers call “AI swarms” capable of “coordinating autonomously, infiltrating communities, and fabricating consensus efficiently,” according to a recent study in Science.
“It’s sort of like a troll farm without actually having to have people anymore,” Wirtschafter said.
Despite these challenges, researchers believe humans can still prevent malicious actors from using AI to destabilize society. The Coalition for Content Provenance and Authenticity has developed technical standards that embed metadata in digital content to verify its origin and any edits.
Several major platforms – LinkedIn, Pinterest, TikTok, and YouTube – have committed to labeling AI-generated content. However, recent testing revealed inconsistent implementation. The most diligent platforms labeled only 67% of AI content, while Instagram labeled just 14% of fake images.
Meta’s oversight board has expressed concern about inconsistent implementation of these standards, even on content generated by the company’s own AI tools.
Gregory attributes this inconsistency to “a failure of political will at the senior levels” of major technology companies. He maintains optimism, stating, “We don’t need to give up on the ability to discern what is real from synthetic. But we do need to act fast.”
Fact Checker
Verify the accuracy of this article using The Disinformation Commission analysis and real-time sources.
8 Comments
While I appreciate the artistic and technological capabilities behind deepfakes, their use for political propaganda is extremely concerning. We must remain vigilant and find ways to empower the public to discern fact from fiction online.
I agree. Fostering digital literacy and critical thinking skills will be crucial as these technologies become more advanced and ubiquitous. Maintaining a well-informed citizenry is essential for the health of our democracy.
This is a troubling trend that requires a multi-faceted response. In addition to technological solutions, we need robust media literacy education to help the public develop the skills to spot manipulated content. Protecting the integrity of our political discourse should be a top priority.
Deepfakes are a complex issue – they can have legitimate uses but also enable malicious misinformation. I’m curious to hear more about the specific policy and technological solutions being explored to address this challenge effectively.
That’s a great point. Policymakers and tech companies will need to collaborate closely to develop robust detection methods and content moderation frameworks that balance free speech with combating disinformation.
The proliferation of political deepfakes is a worrying development that undermines trust in our democratic institutions. I hope researchers and regulators can find ways to better detect and limit the spread of these synthetic media manipulations.
This is a concerning trend that blurs the lines between reality and propaganda. While deepfakes can be used for creative purposes, their misuse for political gain is troubling. We need stronger safeguards and education to help the public discern authentic content from artificial.
The growth of political deepfakes is deeply unsettling. While the technology is impressive, its exploitation for propaganda purposes undermines our ability to have informed, fact-based debates. I hope policymakers and tech companies can work together to address this challenge effectively.