“OpenAI just dropped Sora — a social app that lets you bring your friends into videos with AI magic!”
OpenAI today unveiled Sora, a social video platform powered by its new Sora 2 model, which enables users to generate, remix, and share AI-generated short videos — and even insert themselves or their friends into those videos via an AI “cameo” feature.
What
is Sora & what it can do
- The Sora app is
currently available on iOS (invite-only) in the U.S. and Canada.
- It is built around
the Sora 2 video generation model, which is designed to produce
more physically consistent, realistic motion and synchronized audio than
prior video models.
- In Sora, users can:
·
Start from a text prompt or upload
an image to generate videos in different styles (cinematic, anime, surreal,
photorealistic).
·
Use “cameos” that let you drop your
own likeness (after a one-time video/audio capture) into AI scenes. You can
also permit your friends to use your likeness.
·
Remix or re-generate videos created by
others, creating collaborative and iterative content.
· Browse
a vertical “feed” much like TikTok, with algorithmic recommendations.
Why
this matters — and potential opportunities
- With Sora, OpenAI is
attempting to move AI video generation from a niche research demo into a social,
interactive experience
- The cameo feature is
novel in that it gives users control over their digital likeness —
you can allow or revoke access to others who want to insert you into
generated videos.
- By making the app
invite-based and region-limited initially, OpenAI may be trying to slow-roll
adoption and monitor moderation, misuse, and content quality.
- It positions Sora as
a competitor to short-form video platforms (TikTok, Reels, Shorts), but
with a twist: all content is AI-generated, not user-uploaded.
Risks,
challenges & controversies
- Deepfake / misuse
concerns: Because Sora allows insertion of
likenesses (with consent), critics warn about potential misuse, identity
abuse, or unauthorized deepfakes.
- Copyright &
content sourcing: OpenAI has adopted a default “use
unless opt-out” approach for copyrighted content in video generation,
which has already drawn criticism from rights holders.
- Content moderation
issues: Ensuring harmful, extremist, or
illegal videos are blocked is a major challenge at scale
- Authenticity fatigue:
As Sora’s content saturates feeds, the boundary between “real” and
“AI-made” may blur, raising trust and misinformation concerns.
- Bias and
representation: As with many generative models,
Sora may reflect biases from training data (e.g. gender, race) in how it
portrays people or scenarios.
More link: https://www.indiatoday.in/technology/news/story/openai-launches-new-social-app-sora-lets-you-add-friends-to-videos-using-ai-2796009-2025-10-01