Every day will soon become April Fool's Day

For journalists, April Fool’s Day is the worst day on the Internet. For jokesters, it’s a gold mine of pranks, where “news stories” and “product announcements” trick people into thinking they’re the real deal. Innocently enough it’s all for the laughs, and journalists come prepared –?begrudgingly – with an extra dose of skepticism.

But as generative AI gets better at creating hyper-realistic imagery, sounds and voices, April Fool’s Day will soon repeat itself every day of the year. For many users, the ability to make mind-boggling AI creations is an exciting new feature, not a bug. We’ve played with AI face-altering features for years on apps like Snapchat and embraced photo-bending features like Google Pixel’s “Magic Editor.” Those were just the tip of the iceberg. Driven by user demand, AI manipulation features are getting built into media creation tools everywhere.

For most people, posting AI-altered or synthetic media is not a malicious act. It’s simply entertainment. It gets the ?? and drives engagement. For others it’s creative expression. Soon, social media feeds will be chock full of manipulated media. “Is it real or an AI?” will become one of the most common questions posed on the internet.

This is an easy prediction to make. AI detection tools and watermarking techniques are an imperfect science, and middle schoolers often know how to beat them. Provenance initiatives like C2PA hold much more promise, but they have a long ways to go before they scale. Even a sloppily-Photoshopped photo of Kate Middleton fooled news agencies –?at least for a few hours.

“When you show people deepfakes and generative AI, a lot of times they come out of the experiment saying, ‘I just don’t trust anything anymore,” said researcher David Rand, a professor at MIT.

And that’s the crux of the problem. While creators have fun, journalists will be pulling out their hair. This may be the last April Fool’s Day that’s actually constrained to a single day.