Deadly Magic: The Serious Consequences of Deepfakes

AI-generated images and videos were once comically flawed. The flaws were so obvious that they became jokes: the mouth didn’t quite match the words, the hands looked strange, they looked comically cartoonish, and the risk of someone mistaking it for reality felt far away. Now, you often can’t tell at all. Emerging models are producing visuals and clips that are nearly indistinguishable from real footage, blurring the line between fact and fiction. This has become more than just a technological milestone; it is an inflection point for society with implications beyond entertainment, into security, trust, and even the battlefield.

Deepfakes started as joke videos and silly pranks. But their more nefarious side quickly became the spotlight. After years of rapid improvement, realistic AI-generated images and videos can be easily made by anyone with a laptop, creating convincing videos of a person saying something they never said. This changes something fundamental. Can we believe what we see, what we all believe to be fundamentally true: our eyes? In the past, when video evidence surfaced online, it meant something. Now it might be fabricated. A fake apology. A fake crime. A fake political statement. The barrier to manipulation has collapsed. For everyday people, this already has a human toll. Students, influencers, and public figures have had their likenesses stolen to create non-consensual or misleading content. Reputations can be destroyed overnight by something that never happened. But the stakes don’t stop at personal harm.

Recent events show that deepfake-style manipulation isn’t just limited to anonymous trolls or foreign actors — it's being used in mainstream political discourse. In Minnesota, after civil rights activist Nekima Levy Armstrong was arrested during a protest against increased ICE activities, an official White House social media account circulated a digitally altered image of her appearing to cry. The image exaggerated an emotional reaction she did not display. According to The New York Times, who interviewed her, “the exaggerated features and the darkened skin… reminded her of when the bodies of enslaved people were left disfigured to deter uprisings on plantations, or during Jim Crow when racist propaganda would depict Black people as caricatures.” When criticized, the image was brushed off as a “meme,” which highlights how casually altered media is now treated, even when it shaped public opinion about real people and events.

It comes as no surprise that AI has been integrated into processing intelligence and analyzing footage for the military—cutting-edge technology has always been utilized for military purposes. What is revolutionary is information and psychological warfare. If a single altered image can reshape the narrative around a protest, imagine what a convincing fake video could do during an active conflict. The damage a fabricated rumor, “public” outcry, or “leaked information” could do in days, a deepfake could do in minutes. While humans may not always believe what they hear, they tend to trust what they see with their own eyes. Deepfakes can be used as digital weapons: fake surrender videos, fabricated atrocities, or forged announcements from military leaders. A convincing fake could spark panic, lower morale, or manipulate international opinion before anyone has time to verify it. 

Deepfakes having widespread consequences isn’t a hypothetical. Russia, in its war in Ukraine, has already utilized this novel tool to create videos on social media that depict Ukrainian soldiers appearing to weep and surrender on the front lines. To the average viewer, they look real, mirroring many of the videos that have emerged from the region during these years of conflict. Few have signs of manipulation, and in a world where social media companies profit from the spread of misinformation, this is only adding fuel to the fire. The speed at which these images spread makes correction almost irrelevant. You don’t need bombs to destabilize a country anymore. You just need believable lies that spread faster than the truth. In this new landscape, national security isn’t just about tanks, missiles, or nukes—it’s about who controls information and who can fake it best.

This technology is not going away and will only grow more sophisticated and powerful. That reality forces a choice: either we adapt, or we accept a world where trust becomes optional. This creates a challenge and an opportunity. It has the potential to do much good for the world, but it can cause harm in the wrong hands or even unintentionally. This is not the first technology to present such combinations of potential good and evil, and it will not be the last. But what feels different now is how directly this technology targets our ability to agree on what is real. As technology capabilities accelerate, we must grow adept at navigating the changes it brings to the world and our lives. If we don’t, the cost won’t just be confusion—it could be real human lives.