Is Visual Evidence dying?

Photo via Shuttertalk

***

When the Department of Justice released a slew of Epstein-related documents last Friday, the outrage was palpable. The emails, photos, and testimonies found in the files left public officials scrambling, with hundreds of TikToks and Instagram posts only amplifying the news. Last week, an Instagram reel of protestors heckling NYC mayor, Zohran Mamadani, outside his Upper East Side apartment amassed an impressive 7 million views after his mother was mentioned in the files. The initial photo that sparked this outrage features a young Zohran with his mother, Jeffrey Epstein, and several affiliates, including Bill Clinton and Ghislaine Maxwell. Understandably, New Yorkers were left frustrated and confused. There’s just one caveat: the photo isn’t real. 

The now-viral image was generated using an AI imagery platform and distributed through an X meme account that claimed to create “high-quality AI videos and memes.” Since then, the image has circulated on nearly every social media platform, including TikTok, Instagram, and Facebook. Shortly after the image was debunked, Mamdani publicly stated, “At a personal level, it is tough to see images that you know to be fake, that are patently photoshopped and AI-generated, and yet can reach across the entirety of the world in an era of misinformation.”

A similar misinformation incident occurred on X a few weeks prior, this time at the hands of the official White House account. When Minnesotan lawyer Nekima Levy Armstrong was arrested for peacefully protesting a pastor’s connection to Immigration and Customs Enforcement (ICE), she maintained a calm demeanor. When her run-in with the law was shared on the official White House X account, the image was altered to portray her in hysterics, allegedly using a combination of AI image platforms. “They couldn’t break me by arresting me,” said Ms. Armstrong in an interview with the New York Times, “so they doctored an image to show the world a false iteration of that time to make me look weak.” 

These are not isolated scenarios but rather parts of a growing trend in which AI-generated imagery has begun to supersede legitimate visual evidence. The persuasive nature and rapid expansion of artificial images are now blurring the boundary between visual expectations and actual reality. For rationalists, who rely on pragmatic reasoning, this is alarming. For surrealists, who honor the absurd, this is an inevitable future. 

Transferring the 20th-century surrealist movement that relied on irrationality and abnormal visuals to challenge social norms to a more modern context, we see this same philosophy replicated with the uncanny visuals of AI imagery. Philip Toledano of the Washington Post has dubbed this new era of documentation as “historical surrealism.” While the original surrealist movement produced iconic works and embedded several artists into history, the output of its modern counterpart is more complex. AI redefines surrealism with unreliable visuals that are easily weaponized, as seen with Mamdani and Armstrong. It relies on uncertainty and anxiety to make the viewer stop in their tracks and second-guess what is and isn’t real. 

Traditionally, events and people are visually chronicled through photography, specifically photojournalism. This method relies on impartiality and honesty to communicate with others, reach mass audiences, and compose larger visual narratives. Every end of American history—from political scandals to the peaks of wartime—is preserved through the journalistic honesty and objectivity of photography. Yet, as AI’s historical surrealist lens takes center stage, integrity is getting superseded by absurdity. 

There is an undeniable truth that comes with looking at a photograph. The viewer is aware, whether consciously or not, that what they are looking at is rooted in reality. After all, someone had to be there to take the photo. As AI not only replicates but also dominates photography trustworthiness, this awareness is threatened. 

Suddenly, anything is possible, and as the distinction between artificial imagery and legitimate photographs blurs, it’s difficult for viewers to believe what they see. Institutional authority is already in rapid decline, with only 17% of Americans stating they trust the U.S. government. When figures who are actively winning back this trust, such as Mamdani, are only one ChatGPT prompt away from political disgrace, governmental distrust is accelerated. If we can’t believe what we see, what do we have left? 

There are currently several ongoing discussions regarding the ethics and potential regulation of AI imagery. The U.K. is threatening to ban Grok, Elon Musk’s latest AI model, after the platform was recently used to create sexualized deepfakes of nonconsenting individuals. The leveraging of AI against Mamdani, in both recent weeks and during the 2025 NYC mayoral campaign, has led New York Governor Kathy Hochul to propose a ban on AI use in political campaigns. While there is technically federal regulation of AI imagery with the 2025 TAKE IT DOWN Act, the past month has proven that loopholes and persistence against ordinance will continue to challenge AI image regulation. Visual culture is constantly evolving within broader social contexts, and learning to critically engage with these developments may be the only way through the new age of historical surrealism.

***

This article was edited by Sofia Downes.