One of the more unsettling things about AI is how blisteringly fast the technology develops. A few months ago, it seemed like useful AI-generated video tools were still out of reach. Well, now that’s over. OpenAI just released a new text-to-video generator called Sora. In a word, it’s stunning.
“Sora is able to generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background,” said OpenAI in a blog post. “The model understands not only what the user has asked for in the prompt, but also how those things exist in the physical world.”
Here’s how Sora’s video are going to exist in the world: in the weirdest and most uncanny way imaginable.
For now, OpenAI is running final tests on Sora to make sure it’s ready for public consumption, and only a handful of select users have access to the tool. But based on what the company and early users have shared, it’s clear the tool is about as capable as the best image generators on the market. There are weird, tell-tale signs you’re looking at AI hallucinations, but when Sora is working well, it can be hard to distinguish the AI’s work from videos captured in the real world.
In the near future, the consequences will be Earth-shaking. The world will fill with deepfakes. People will lose their jobs. Artists will have unprecedented and exciting new opportunities. There will be questions about where OpenAI got the training data. You’ve heard this all before, but video is an entirely different animal from text and images. With widely available AI-video generators at people’s fingertips, we’re about to enter a new era.
The web is slowly filling up with people’s Sora creations. Click through the slideshow to get a preview of what your life is going to look like in the near future. If you’re on mobile, just scroll down.