On October 3, 2025, OpenAI took a bold leap into the social media landscape with the launch of Sora, an app that’s already stirring up excitement, curiosity, and more than a little concern. Built around the company’s upgraded Sora 2 AI video generator, this TikTok-style platform invites users to create, share, and endlessly scroll through short-form videos—except, in a twist that feels uniquely 2025, every clip is generated entirely by artificial intelligence.
The premise is as simple as it is radical: Sora users type a prompt, and the AI responds with a highly realistic, 10-second video. According to NPR, early users have already filled the app’s feed with everything from a fascist SpongeBob SquarePants and a dog behind the wheel to Jesus playing Minecraft. If you can imagine it, Sora can likely create it—provided you don’t run afoul of its moderation filters.
OpenAI’s Sora isn’t just another AI tool; it’s a full-fledged social media experience. The app’s interface, as described by CNET, will feel familiar to anyone who’s spent time on TikTok or Instagram Reels: an algorithmically curated “For You” feed, options to like, comment, share, and follow, and even the ability to filter videos by your mood. But lurking behind that user-friendly façade is a technology that’s both dazzling and a little unsettling.
One of Sora’s most talked-about features is its “cameos.” When users sign up, they can upload their own likeness—essentially giving permission for the AI to insert them into any number of scenarios, real or imagined. “Every Sora user can upload their cameo when they make an account and choose whether to let other people use their likeness in AI videos,” CNET reports. It’s a feature that’s already been put to the test with OpenAI CEO Sam Altman, who’s quickly become one of the most cameoed figures on the platform. Within minutes of scrolling, users have reported seeing videos of Altman in wildly fictional situations: getting arrested for stealing GPUs, singing ballads about inference costs, or even stuffing his nose with tissues while begging for likes. As CNET’s reviewer put it, “Altman, presumably, has never done any of these things in real life. But the quality of the Sora 2 AI videos is so realistic that you wouldn’t know otherwise.”
That realism is no accident. The Sora 2 model boasts significant improvements over its predecessor, particularly in dialogue generation and complex prompt handling. Unlike Google’s Veo 3, which often requires detailed scripts, Sora can spin up convincing conversations or even write song lyrics from a simple phrase. The trade-off, according to CNET, is that videos take a bit longer to generate—between two and five minutes—but the results are a far cry from the low-quality “AI slop” that’s become infamous elsewhere on the internet. “On the whole, though, the audio is clear, the text is error-free, and the videos are decent and deceptively human-like. It’s scarily realistic,” the reviewer notes.
Still, the potential for abuse is hard to ignore. Sora’s moderation filters block attempts to create videos featuring certain celebrities—Taylor Swift, for example—but users have already found ways to skirt these restrictions. One video, flagged by CNET, showed Altman in a field with Pikachu, joking, “I hope Nintendo doesn’t sue us.” The app tries to mitigate risks by including visible watermarks and metadata tags on downloaded videos, clearly disclaiming their AI origins. But as the reviewer warns, “in a fragmented world of declining trust, Sora’s AI deepfakes are too easy to make and too easy to abuse.”
For all its technical prowess, Sora is not without its frustrations. The app currently lacks advanced video editing tools—no trimming, no closed captions, and no custom post captions. If you want to make changes, you have to regenerate the entire video, which can be a tedious process. “I even found myself frustrated that Sora didn’t have better video editing tools when I was creating cameos,” CNET’s reviewer admits. At launch, access to Sora was restricted to those with an invite code, adding a sense of exclusivity—and, perhaps, a bit of viral buzz.
OpenAI’s stated goal with Sora is to “bolster human connection.” Yet, as users and critics alike have pointed out, the experience can feel more isolating than unifying. The endless scroll of AI-generated content is highly addictive, but it’s also vaguely disconcerting. As the CNET author reflected, “I can’t say I felt closer to my fellow man after my initial intrigue wore off. Sora videos contain a visible watermark when you download them, along with metadata tags that disclaim their AI origins. But in a fragmented world of declining trust, Sora’s AI deepfakes are too easy to make and too easy to abuse.”
The timing of Sora’s release is no accident. Just days before, Meta rolled out its own AI video tool as part of the Meta AI platform, signaling a race among tech giants to dominate this new frontier of content creation. NBC News highlighted Sora’s launch in a recent segment, noting the growing blurriness between real and fake media as AI-generated videos become more sophisticated. The implications are profound—not just for the entertainment value of surreal clips like “Jesus playing Minecraft,” but for the broader question of what (and whom) we can trust online.
Researchers have been both wowed and worried by Sora’s capabilities. NPR found that the app could “easily generate very realistic videos, including of real individuals (with their permission). The early results are both wowing and worrying researchers.” With the ability to effortlessly create deepfakes, the stakes for misinformation, privacy, and digital consent have never been higher.
Yet, for all the hand-wringing, there’s a sense that Sora is simply the logical next step in our relationship with technology. We’ve already seen how AI tools can speed up creative workflows, generate bespoke content, and even help us fight insurance denials or avoid doctor burnout, as NBC News has reported in other contexts. Sora just takes those possibilities and packages them in a way that’s instantly accessible—and, perhaps, a little too easy to lose yourself in.
As the dust settles on Sora’s launch, one thing is clear: this app is both a marvel of modern technology and a mirror reflecting our deepest anxieties about the digital age. Whether it will truly “bolster human connection” or simply flood the internet with a new kind of AI-generated slop remains to be seen. But for now, Sora is exactly what an AI social media app in 2025 would be: mesmerizing, unsettling, and impossible to ignore.
