Sora 2 from the team behind ChatGPT enables anyone to produce incredibly realistic social media videos, including some great car clips
by Chris Chilton
1 hour ago
- OpenAI’s Sora 2 video-generator launched on September 30.
- It’s more realistic than the original and features native sounds.
- Users can put different characters into the vids, including them.
If your social media feeds are anything like ours, there’s a good chance a big chunk of the content you see every day has people doing funny, dangerous, or simply exciting things with cars. Every day brings another clip of someone pushing the limits, sometimes intentionally, sometimes not.
But what if you didn’t have to hope that something car-related and laugh-out-loud funny was going to happen somewhere in the world while someone with a smartphone was around to capture and upload it? What if you could just make it yourself in AI?
Related: The World’s Most Famous Italian Design House Just Let AI Take The Pencil
That’s what’s happening right now as thousands of people are exploring the possibilities of Open AI’s Sora 2. The TikTok-like app lets users generate videos that can be watched on the original platform, or shared to others like TikTok or X.
Reality Gets a Rewrite
Sora was launched last year but the second-generation version, which has only been available since September 30, makes a big leap forward. It has better physics, synchronized audio and, more importantly, it – mostly – looks entirely believable.
Which is absurd when the clips we’ve added below show Scooby and Shaggy and SpongeBob Squarepants getting pulled over by the police. They’re obviously not real, they can’t be.
Still, your brain can easily accept them as real in the same way that we get sucked into watching a Pixar animation and quickly stop thinking about what we’re seeing as lines of computer code.
Brilliant, Terrifying, or Both?
The clips that are less ridiculous, but still outlandish, like Michael Jackson stealing somebody’s dinner in a fast food restaurant, are brilliant and terrifying at the same time. They highlight just how fast the technology is evolving and how easily it can blur the line between reality and fabrication.
Also: Americans Embrace AI Tech In Their Cars But Some Features Drive Them Crazy
Going forward, we might have to be careful about believing what we see on our screens, whether it’s a Mustang wrecking in spectacular fashion after leaving a car meet (even if SpongeBob isn’t driving), or a clip of a prominent politician saying or doing something that hurts his reputation.
When Fun Turns Problematic
And because this new breed of AI lets you put yourself and your friends into videos, there’s the potential for it to go badly wrong, even if the majority of users just treat it as a bit of fun. Imagine going for a big promotion at work in 10 years time and losing out because someone has uploaded a picture of you failing a sobriety test and assaulting a cop.
More: This Unstoppable AI Patrol Car Hunts 1,500 Parking Violators Every Hour
Soro 2 looks like real fun and watching Stephen Hawking dropping in off a skateboard ramp in his chair was mildly amusing, though soon after launch OpenAI agreed to give concerned copyright holders more control over how and if their characters could be used. But the future it points to is downright scary.
Image Sora/YouTube
#Traffic #Stop #Videos #Equal #Parts #Hilarious #Terrifying
