If your feed isn’t already filled with AI-generated videos, it’s just a matter of time.
Meta and OpenAI will make sure of that. Meta recently announced its Endless Vibes, made up entirely of AI-generated content: cats, dogs, and spots. And that’s just in Mark Zuckerberg’s initial video post about it.
OpenAI’s new Sora app offers a different flavor of trash. Like TikTok, Sora has a For You page to scroll through content vertically. But the scariest part about Sora is how real he looks. One feature, called Cameo, allows users to make videos of themselves, their friends, and any public profiles that grant them access. This means that videos of Sam Altman hanging out with Charizard or interrogating Pikachu are circulating on social media. And, of course, videos of Jake Paul are also starting to circulate.
It’s just the beginning and technology is only getting better. To help us navigate it, we spoke with Hayden Field, senior AI reporter at The Verge. field and Today, explained Co-host Sean Rameswaram discusses why these tech giants are doubling down on AI videos, what to do with them, and we even got fooled by one.
Below is an excerpt from the conversation, edited for length and clarity. There’s a lot more in the full podcast, so give it a listen Today, explained wherever you get podcasts, including Apple Podcasts, Pandora, and Spotify.
What is Mark Zuckerberg trying to do with Vibes?
That’s the million dollar question. These companies, especially Meta right now, really want us to continue consuming AI-generated content and they really want to keep us on the platform.
I think it’s really just Zuckerberg trying to make AI a bigger part of the average person’s life and routine, getting people used to it and also putting a sign in the ground saying, “Hey, look, this is where the technology is right now. It’s a lot better than when we saw Will Smith eating spaghetti.”
How could he improve so much and so quickly? Because yes, this is not Will Smith eating spaghetti.
The AI now trains itself most of the time. You can improve and train to improve. One of the big things that stands in your way is computing. And all of these companies are building data centers and doing new business every day. They’re really working on getting more computing, so they can push the technology even further.
Let’s talk about what OpenAI is doing. They just released something called Sora 2. What is Sora?
Sora is their new app and it is basically an AI generated endless scrolling video social media app. So in a way, you can consider it as an AI-generated TikTok. But the craziest part, honestly, is that you can make videos of yourself and your friends too, if they give you permission. It’s called Cameo and you record your own face moving from side to side. You record your voice speaking a sequence of numbers and then the technology can parody you doing any number of things you want.
That’s why it’s so different from Meta’s Vibes and why it feels different when you scroll through it. You are watching videos of real people and they look real. I was browsing and I saw Sam Altman drinking a giant juice box or a bunch of other things. It seems like it really is Sam Altman or it seems like it really is Jake Paul.
How do we know if what we are seeing is real or not in this era in which it is increasingly difficult to discern?
These tips that I am about to give you are not infallible, but they will help you a little. If you look at something long enough, you’ll probably find one of the telltale signs that something is AI-generated.
“Taylor Swift, actually, some of the promos for her new album apparently had a Ferris wheel in the background and the spokes became blurry as it moved.”
One of them is inconsistent lighting. Sometimes it is difficult for AI to correctly pick up the vibrations of a place. If there are a lot of lamps (maybe it’s really dark in a corner, maybe it doesn’t have the realistic quality of sunlight), that might be something you can spot. Another thing is unnatural facial expressions that just don’t look quite right. Maybe someone is smiling too much or crying with their eyes too open. Another is retouched skin, skin that looks too perfect. And finally, background details that can disappear or transform as the video progresses. This is a big one.
In fact, Taylor Swift: Some of the promos for her new album apparently had a Ferris wheel in the background and the spokes became blurry as it moved.
Is there anything else we should be looking for?
I just wish we had more rules about these things and how they could be disclosed. For example, OpenAI has a protection: every video you download from Sora has a watermark or at least most of the videos. Some professional users can download one without watermark.
Oh cool, if you pay them money you might lose the watermark. Very pretty.
But the other thing is that I’ve seen a lot of YouTube tutorials that say, “Here’s how to remove Sora’s watermark.”
Do companies like OpenAI or Meta care if we can know if this is real or not? Or is that exactly what they want?
They say they care. So I guess that’s all we can say now. But it’s difficult because, by the very nature of a technology like this, it will be misused. So you just have to see if you can curb that misuse as much as possible, which is what they’re trying to do. But we’ll have to wait and see how successful they are at that. And right now, if history is any guide, I’m a little worried.