Stop Worrying About Deepfakes

How are we to know what is real, on the internet? When a faked photo of an explosion near the Pentagon briefly spooked Wall Street this spring, it was the first major case of a viral AI-generated image moving the market. The hoax wilted under scrutiny, but to many, it seemed like a harbinger of far worse to come.

And yet, when Walter Scheirer, a computer scientist and media forensics expert at the University of Notre Dame, sent his students to scour the internet for examples of AI-doctored videos, what they brought back surprised him. It was, he says, “memes all the way down.”

So far, the majority of convincing deepfakes seem to be those engineered to generate a knowing chuckle rather than an economic collapse: fake Tom Cruise talking about hand-washing; Nicholas Cage in every movie ever; Joe Biden, Donald Trump, and Barack Obama playing Call of Duty together. The internet, Scheirer concluded, is indeed overflowing with fake content, but the vast majority of it seems aimed at the creation of connection—rather than destruction.

Scheirer and the internet grew up in tandem. An older millennial, he describes his first computer as “primitive”—a Texas Instruments box that hooked up to the family TV. In middle school, he began hanging out in Internet Relay Chat hacker channels, rubbing virtual shoulders with the. Even then, he says, hackers loved twisting science facts and fiction together, for the craft of it.