The Future of Deepfakes and the Decline of Reality (With Hany Farid)
404 Media Emanuel Maiberg
404 Media talked to Hany Farid about deepfakes and where they’re headed. He’s the guy people call when AI images get too convincing, and he’s been at this since 2017.
Based on reporting by 404 Media, Emanuel Maiberg — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
404 Media is back on the deepfake beat with Hany Farid, a name that keeps coming up any time synthetic images start looking a little too real. Farid co-founded GetReal, which focuses on spotting deepfakes and other AI-made images, and he also built PhotoDNA, the hashing system used to help platforms find and remove some of the worst material online.
That mix makes him a useful person to hear from now, because this isn’t a new panic. 404 Media says it started talking with Farid regularly in 2017, when Sam first reported on deepfakes, and the conversation has kept shifting as the technology changed and reality got harder to trust at a glance.
The latest podcast episode is framed as a look back and a look ahead: what has changed, what still holds up, and where this is all going next. That’s the right lens. Deepfakes stopped being a novelty a while ago. The more interesting question now is how people, companies, and platforms keep up when seeing is no longer enough.
404 Media is also using the episode to push its subscriber offering, with bonus content, extra comment-response episodes, early access to interview series, and a subscribers-only feed. But the main draw here is still Farid himself: someone who has spent years trying to build better tools for telling real from fake while the internet keeps making that job harder.
My take — AI-written commentary, not fact-checked reporting
The depressing part isn’t that deepfakes exist. It’s that the people best placed to fight them are now also the people explaining why the old rules stopped working. That’s the whole internet story in miniature: trust gets automated, then suspicion gets automated, and everyone calls it progress because the demo looked smooth.
Read more about this at: 404 Media