Sora2: Make Free Deepfakes of Yourself

December 2025

I was having so much fun making deepfakes. Here I was, galloping up a country road on horseback, hitting a perfect tennis serve, joining ICE patrolling the streets of Chicago, wearing a frog costume while participating in the rebellion in Portland.

The new, free Sora2 app first has you make a “cameo” of yourself, which entails looking at your phone’s camera, reading aloud three two-digit numbers that appear on the screen, then looking left and right. The app then uses AI to recreate your appearance and voice. You can then insert yourself into any imaginable situation.

I’m a wannabe tennis player who has been striving for years to get the correct form on my serve. This was my chance! The 10-second deepfake begins with me saying, “I’ve finally mastered the serve.” And then shows me hitting a perfect serve.

It was oddly satisfying, as if that vicarious experience somehow helps to make up for my failure. The next time on the court, I showed it to my partner before we began play. He said, “That looks good.” Then I admitted it was fake. Sigh.

Sora makes 10- to 15-second videos. You specify the details and tell it what you want your image to do and say. I combined my first five videos into a 50-second movie and posted it on Facebook. My friends there were surprised by the novelty. A couple were eager to try Sora. But it didn’t get nearly as many reactions as my nature videos.

Then after a couple days, I got an alert from Facebook saying that my little video, which I had made public, had reached 200 views. Yes! It was going viral. The alert even encouraged me to make another.

I think, though, that the novelty of seeing yourself in various situations wears off, both for me and the friends on whom I imposed. I thought maybe I could do something more creative. And somehow ended up with a crude theme: needing to find a restroom.

I had Sora create a video of me walking around on the moon, with the lunar landing module in the background. I can then be heard saying, “Mission control, can you tell me where to find a restroom?” I laughed for a couple days thinking of this and my next video.

In it, I’m standing in front of the Lincoln Memorial, speaking in a stentorian voice to a cheering crowd of 250,000 on the National Mall. “I have a dream, a dream for the United States of America.” Big cheering from the crowd. “But first…I need to find a restroom.”

It’s astonishing what Sora creates based on a simple prompt. In this 10-second video (and in the others), Sora used several angles, including one from behind my head showing the crowd in dramatic fashion.

Next, I’m a scuba diver on the bottom of the ocean. I say to myself, “I sure could use a restroom.” Then as I swim, I see a porta-potty up ahead on the ocean floor.

As with my first collection of Sora shorts, I ended the composite video I made by showing me sitting at a computer and making a bemused comment.

Of course, I then posted on Facebook. Comments included, “OMG that is hilarious!!!!” and “So funny!”. But fewer reactions this time, though again picking up lots of views on Facebook's collection of public videos called reels.

Sora2, from OpenAI, is free, and hugely popular. It first launched for the iPhone and iPad on September 30th and had over 1 million downloads in less than five days—better than ChatGPT’s launch. In early November OpenAI released a version for Android devices.

It’s not only a tool for creation but also a social media service similar to TikTok. Users can share their creations on the platform and even incorporate multiple cameos, allowing them to create videos that feature themselves along with their friends.

Huge downsides, though. Each video creation uses considerably more energy than a ChatGPT query. And Sora is contributing to the flood of “AI slop” on the internet. Plus, OpenAI has controversially allowed videos of deceased people, including scatological videos of Queen Elizabeth II.

In addition, the ubiquity of deepfake videos means that one can no longer trust anything one sees. Fortunately, OpenAI puts a tiny Sora logo on every Sora video, though that can be edited out.

It is so amazing that this powerful tool is free. And so amazing how AI keeps forging ahead. Yet, that too, is problematic. Hundreds of billions of dollars are being invested in AI, but revenue is relatively minuscule. Some economists fear that we’re facing a dangerous bubble like the ones that crashed the economy in 2001 and 2008.

I think I’m done with Sora for now. But it’s been great fun. You can see my videos here.