OpenAI to Launch “Sora 2” Stand-Alone App with TikTok-Like Interface and Content Strategy, Copyright Policy Sparks Debate

Sam Altman, CEO of OpenAI

**September 30 —** According to insiders, OpenAI is preparing to launch the next-generation video-generation model **Sora 2**, along with a stand-alone mobile application. The app’s **functionality and interface closely resemble TikTok**, signaling OpenAI’s ambition to bring AI-generated video creation to the mass market.

More controversially, OpenAI will introduce a new **“opt-out” mechanism** for Sora 2. This policy allows copyrighted material to be used in AI-generated videos **unless rights-holders explicitly request to exclude their content**.

Over the past week, OpenAI has begun notifying major talent agencies and film studios of the new system. This means intellectual-property owners — such as movie studios — must proactively inform OpenAI if they wish to **block their copyrighted material** from appearing in Sora-generated videos. While the new app will not generate identifiable likenesses of real public figures without permission, **fictional characters protected by copyright must go through the “opt-out” process** to avoid being used.

TikTok-Inspired Design: An AI-Native Short-Video Platform

According to documents obtained by *Wired*, the upcoming **“Sora 2 App”** adopts a **vertical video feed** with swipe-based navigation **nearly identical to TikTok’s**. Users can create video clips of up to **10 seconds**, with a recommendation algorithm powering a **“For You” page**. A right-hand action bar enables **likes, comments, and “remix” editing** features.

The app will **not allow uploads** of photos or videos from users’ phones or other apps — all content must be **natively generated by Sora**. It also includes a **portrait-verification feature**: once users verify their likeness, they can appear in their own generated clips, and others may tag or reference that likeness. Regardless of publication status, the verified user will receive a notification whenever their portrait is used.

Internal OpenAI documents show that the app was released for internal testing last week and received **overwhelmingly positive feedback**. Employee engagement was so high that some managers joked it might be **hurting overall productivity**.

Racing for the AI-Video Crown: Copyright and Safety Collide

OpenAI first introduced its **Sora video model** in December 2024, later integrating it into ChatGPT. By launching Sora 2 as a stand-alone app, OpenAI is now **entering direct competition with Meta and Google** in the fast-emerging AI-video arena.

Just last week, **Meta** launched a *Vibes* feature within its Meta AI app for creating and sharing AI-generated short videos, while **Google** announced plans to integrate a customized version of its Veo 3 model into YouTube.

In contrast, **TikTok** remains cautious about AI-generated media. Its latest policy explicitly bans any AI content that is *“misleading on matters of public importance or harmful to individuals.”*

Meanwhile, OpenAI faces multiple **copyright lawsuits**, including a high-profile case brought by *The New York Times*, which alleges that OpenAI used the paper’s copyrighted materials for training. Reports indicate that even now, **Sora 2 occasionally blocks certain video outputs due to copyright filters and content-safety restrictions**.

Beyond intellectual property, OpenAI is also confronting **child-safety concerns**. On September 28, the company rolled out new **parental-control tools**, allowing parental-teen account linkage and announcing an upcoming **age-prediction system** that will automatically redirect users under 18 to a more restricted ChatGPT experience. It remains unclear what **age-restriction measures** the Sora 2 App will implement.

This article was originally published by Tencent Tech, author Wuji, and reprinted by 36Kr with permission.