The manipulation of images and videos is not a new phenomenon. As early as the 1860 presidential campaign, photographer Mathew Brady subtly altered a picture of Abraham Lincoln to make him appear more attractive by enlarging his shirt collar to hide his bony neck and prominent Adam’s apple. After Lincoln’s assassination, artist Thomas Hicks created a memorial portrait by transposing Lincoln’s head onto the more muscular body of John C. Calhoun, a pro-slavery advocate, to make the fallen president appear heroic.
These early manipulations were time-consuming and required specialized expertise. However, the advent of AI-generated ‘deepfakes’ has democratized the creation of fake images and videos, making them easier to produce than ever before. Tools like Midjourney and OpenAI’s DALL-E allow users to generate convincing fakes with just a text prompt, lacking any real-world reference or original source.
The Rise of Deepfakes
The ease of creating deepfakes has raised concerns about their potential to deceive people and spread misinformation on social media. In 2021, a series of computer-generated videos featuring a realistic but fake Tom Cruise fooled millions on TikTok. While these were humorous, the creation of fake, sexually explicit images of celebrities in 2024 was not. X (formerly Twitter) was flooded with AI-generated pornographic images of Taylor Swift, forcing the platform to temporarily block searches for her name.
Deepfakes are particularly suited for political dirty tricks. During the 2023 Chicago mayoral election, an AI-generated audio clip circulated on social media featuring a candidate praising police brutality. In the 2024 New Hampshire presidential primaries, a robocall mimicked Joe Biden’s voice, urging independents to skip voting. The Republican National Committee released an ad featuring deepfake images depicting a dystopian future if Biden were reelected.
The Impact on Society
Experts warn that as deepfakes become more prevalent, they may lead to a ‘deepfake inflection point’ where forged media floods the internet. This could have profound implications, not just by making people more gullible but also by increasing cynicism. OpenAI CEO Sam Altman suggests that the threat will diminish once people get used to deepfakes, but some experts believe the opposite: that the risks will grow as we become accustomed to them.
As deepfakes become commonplace, we may start doubting the veracity of all information presented through media, potentially leading to a world where our default is to take nothing at face value. Law professors Bobby Chesney and Danielle Citron caution that this could have significant political implications, benefiting autocrats and harming democracy. ‘Authoritarian regimes and leaders with authoritarian tendencies benefit when objective truths lose their power,’ they argue.
The Role of Mythmaking
The rise of deepfakes and AI-generated content may signal a shift towards a more mythological understanding of reality, where the line between fact and fiction blurs. The Book of Veles, a purported ancient manuscript revealed in the 20th century, is a historical example of mythmaking. Despite being debunked as a hoax, it gained a following among certain groups.
History and psychology suggest that generative AI will make the virtual feel more authentic than reality. Mythology’s subjective and aesthetic approach to understanding the world, which appeals to emotion and the subconscious, has always maintained a hold on the public mind. In an age overwhelmed with information, myth provides a framework for making sense of chaos.
The spread of conspiracy theories may be better understood as a form of mythmaking, appealing to our desire for meaning and explanation. When evidence seems unreal, strangeness itself can become a criterion for truth, and a paranoid logic takes hold. As the romantic poet John Keats wrote, ‘Beauty is truth.’ In the context of deepfakes and AI-generated content, this phrase takes on a new significance, highlighting the complex relationship between perception, reality, and truth.