OpenAI’s Big Bet: Sora 2 and the Future of AI-Generated Video Feeds

The social media world might be about to change again — and this time, it’s not because of a new dance challenge or a viral meme. OpenAI, the company behind ChatGPT, is stepping into the short-form video game with a brand-new app powered by its latest video model, Sora 2. And if it works the way they hope, it could reshape how we create and consume video content online.

The concept is simple but revolutionary: imagine TikTok, but every single clip you scroll past was generated entirely by AI. No camera. No editing. Just words turned into moving images — fast.

A TikTok for AI Video

This new app mimics the familiar vertical scroll we’ve all gotten used to on TikTok and Reels. But instead of uploading your own videos, you generate them with AI. Users type a prompt, and Sora 2 spits out a short clip — up to about 10 seconds — right there in the feed.

The app even includes a “For You” page that learns from what you like, comment on, or remix, feeding you more content tailored to your tastes. It’s a clever twist on a proven formula — but instead of endless human-made videos, the feed could become an ever-evolving stream of synthetic creativity.

Identity and Safety in an AI World

One of the most interesting — and potentially controversial — features is identity verification. OpenAI says users will be able to verify their likeness, allowing the model to generate clips featuring them. It will also notify people if their likeness is used in generated content, even if that video never goes public. That’s a big deal in the fight against deepfakes, but it also opens new questions around privacy and consent that OpenAI will have to navigate carefully.

The company is also building in guardrails around copyright and safety. Sora 2 already refuses to generate certain types of content due to copyright concerns or harmful material. That’s not surprising, given how closely regulators are watching generative AI and how many lawsuits are already brewing over training data and content misuse.

A New Kind of Content Platform

Early testers inside OpenAI are reportedly hooked — using the app so much that it’s become a bit of a distraction. That’s usually a good sign that a product has potential. But potential isn’t guaranteed success.

For one, OpenAI is entering a competitive and crowded space. Meta is pushing its own AI-generated video product called Vibes, Google is weaving video AI into YouTube, and TikTok is experimenting with stricter rules around AI content. The question isn’t whether people will use AI video tools — they already are — but whether a fully AI-powered social platform can stand on its own.

Why This Matters

If OpenAI pulls this off, it could signal the next major shift in online media. We’re used to feeds filled with videos made by people — influencers, brands, friends, strangers. But what happens when those videos are generated entirely by AI, tailored perfectly to your interests, and produced in seconds?

It’s a future that feels both exciting and unsettling. On one hand, the creative possibilities are endless. Anyone could become a “creator” just by typing ideas. On the other hand, it raises new challenges around truth, identity, consent, and how we define authenticity online.