What to know
- OpenAI's new app closely resembles TikTok with a vertical video feed of AI-generated content only.
- Users can create videos up to 10 seconds using Sora 2's advanced video generation model.
- The app includes identity verification allowing users to appear and be tagged in AI videos.
- No user uploads allowed; videos are entirely AI-generated and come with notification safeguards.
OpenAI is preparing to launch a standalone social app powered by its new video generation AI model, Sora 2. The app will feature a TikTok-like experience with a vertical video feed that users can swipe through continuously. However, unlike TikTok, every video on Sora 2 will be fully generated by AI; the platform does not allow uploading personal photos or videos. The videos are limited to 10 seconds long, shorter than TikTok’s current 10-minute limit, focusing on quick, AI-driven creativity.
How Sora 2 will work for users
Users will interact with a personalized feed similar to TikTok's "For You" page, driven by a recommendation algorithm tailored for AI-generated clips. The app will include a menu on the right side where users can like, comment, or remix videos, fostering engagement and creativity within the AI content ecosystem. Identity verification is a key feature, enabling users to authenticate their likeness, which can then be used in AI-generated videos by themselves or others. Users will also get notified when their likeness is used, even if the video is a draft that’s not published.
Key features of OpenAI's TikTok like app
Here are some of the key features of the upcoming app:
- Videos are capped at about 10 seconds.
- TikTok-like interface: vertical scroll, swipe navigation, engagement buttons.
- No uploads of personal photos or videos—only AI-generated outputs.
- Identity verification lets people enable their likeness for generation and remixing. Notifications appear when someone’s likeness is used, even in draft videos.
This ought to ensure the platform runs on a fully synthetic, AI-native content ecosystem.
Building on the original Sora, released in late 2024, Sora 2 offers improved video generation capabilities, including better physics simulation, more realistic human motion, and integrated audio elements like voiceovers and ambient sounds. These enhancements aim to produce short clips that are visually and aurally coherent, enhancing user experience and creative potential within the app.
What copyright, rights, and ownership concerns could OpenAI face
OpenAI intends to implement an opt-out system for copyrighted works, meaning material can appear in generated videos unless rights holders request exclusion. The tool reportedly will not produce recognizable depictions of public figures without permission.
Critics warn this shifts the burden onto artists and studios to actively monitor and request removals, potentially eroding copyright safeguards.
Here is a gist of everything we know so far:
| # | Feature | Details |
|---|---|---|
| 1 | Video length | Up to 10 seconds per video |
| 2 | Content type | 100% AI-generated, no user uploads |
| 3 | Interface style | Vertical swipeable feed, similar to TikTok |
| 4 | Engagement tools | Like, comment, remix videos |
| 5 | Identity verification | Users can authenticate likeness for AI video use |
| 6 | Notifications | Sent when likeness is used, even in unpublished clips |
| 7 | Copyright safeguards | Requires opt-out by rights holders for copyrighted content |
| 8 | Competitors | TikTok, Meta Vib, Google Veo 3 |
Sora 2 not only challenges traditional video platforms but also represents a significant step in mainstreaming AI-generated content as a social and creative medium. Its success could redefine online content consumption and production, opening new possibilities for AI-driven entertainment and social interaction.
OpenAI’s move into a TikTok-like AI video app is timely given current geopolitical factors and growing interest in AI creativity, positioning the company as a key innovator in the emerging AI social media landscape.
Discussion