AI videos: Sora 2, TikTok-style, identity verification
OpenAI is apparently preparing a TikTok-like app that uses exclusively AI-generated videos. Reports point to an evolution of Sora, OpenAI's video-generation model. This app is supposed to integrate features such as identity verification and copyright mechanisms. The current status and possible implications are discussed below.
Quelle: Custom display
OpenAI Sora App
Sora is OpenAI's model for video generation that can generate short clips from text, images, or videos. OpenAI describes Sora as a system that simulates movement of scenes. Currently Sora is available for ChatGPT Plus, Business and Pro. The rumors focus on a possible 'Sora 2' generation that would drive a standalone TikTok-style app. OpenAI Sora was originally introduced as a text-to-video model and documented this in subsequent posts. OpenAI
According to a report by Wired OpenAI is preparing a standalone social app for AI videos. It is supposed to offer a vertical feed, swipe-to-scroll and a For-You algorithm, but exclusively AI-generated content. Users should be able to generate clips up to 10 seconds in the app context; uploads from the camera roll are not planned. An identity check should allow users to confirm their own likeness. Verified individuals could use and be marked with their persona, including notification even for drafts. The app was reportedly started internally last week and received strong employee feedback; OpenAI does not comment on this.
Quelle: geeky-gadgets.com
OpenAI Sora: Text-to-Video AI.
Why an own app? A 'pure AI' video platform lowers the barrier to mass visibility of Sora's everyday experiments, similar to how ChatGPT made text popular. Wired
It has been reported that Wired describes the feed, the 10-second limit, no camera uploads, identity verification with notifications, and the internal test launch. Wired It is unclear whether the name 'Sora 2' and the exact app roadmap are final, as OpenAI has not commented according to Wired. Likewise, exact launch countries, age limits and the scope of the terms of service are unclear since there are no official details in the reports. Wired
If the app comes, you can experience AI video as your own social medium: short, remixable clips, potentially with your verified likeness. This offers opportunities for storytelling, bite-sized learning, and campaigns. Wired
Quelle: YouTube
Short video: Official Sora teaser provides context for how text-to-video at OpenAI is fundamentally conceived. YouTube
Market landscape & competition
OpenAI's efforts to launch its own AI-video app take place in a dynamic market environment. Meta is currently launching 'Vibes', an AI video discovery space that emphasizes remix as a principle. Meta In parallel, YouTube integrates Veo 3 from Google DeepMind into Shorts, including generative audio features. YouTube These developments indicate an emerging AI video arms race.
Quelle: theaidb.com
Sora: The future of AI-powered video creation.
TikTok is also tightening its AI rules and bans content that is misleading or harms individuals when it concerns 'matters of public importance'. This tightens the boundary for deepfakes. The claim 'TikTok loosens AI rules' is misleading; on the contrary, the known changes emphasize prohibitions on misleading AI content with public relevance or personal harm. TechCrunch The claim 'TikTok loosens AI rules' is misleading; on the contrary, the known changes emphasize prohibitions on misleading AI content with public relevance or personal harm.
Media reports paint a picture of a strategic response to the short-video ecology: Meta is launching 'Vibes' for AI videos, while YouTube Veo 3 docks. Meta YouTube
Quelle: YouTube
Short video: Veo-3 demo shows how competing AI video systems generate scenes with sound – useful for comparisons. YouTube
Legal aspects
In parallel with the app developments, Reuters reports on a WSJ-Papier: A new Sora procedure could enable copyrighted content, provided rights holders do not actively object (opt-out); public figures would be excluded. An opt-out approach for copyright would enlarge the repertoire for everyday prompts, but would also create new fault lines with rights holders. Reuters
Quelle: weel.co.jp
Sora in the context of TikTok-like AI videos.
Reuters cites WSJ reports on the planned opt-out for copyrighted material, without generating images of identifiable people without consent. Reuters
Industry commentary emphasizes tensions around copyright and personality rights; Reuters points to an opt-out model as the new minimum standard for rights control. Skeptical voices warn about platforms where mass amounts of synthetic clips circulate and moderation becomes a constant task; TikTok's rule update provides the counterpoint. Reuters Skeptical voices warn about platforms where mass amounts of synthetic clips circulate and moderation becomes a constant task; TikTok's rule update provides the counterpoint. TechCrunch
Open questions remain: How will OpenAI legally secure identity usage – is verification plus notification enough, or are granular consents required for each prompt? Wired Will opt-out indeed become the default for copyrighted material, and how efficient is that for rights holders to implement? Reuters
Impacts & Opportunities
At the same time, due diligence requirements are increasing: pay attention to provenance labeling, context and rights, especially when the platform uses opt-out for protected material. Reuters Official provider resources and transparent changelogs help with orientation; for cross-checks you can use YouTube/Meta product blogs to understand feature parity. YouTube Meta
The hints converge on a short, AI-exclusive video app centered around Sora 2: a format that combines usability, remix capability, and identity politics – and thus creates new creative options, but also new auditing and protection obligations. Until OpenAI officially confirms it, much remains provisional; for you, it's worth keeping an eye on rule changes of the major platforms and OpenAI's rights policy. Wired Reuters