OpenAI has taken an ambitious step into the social media market with the launch of ‘Sora’, its new app designed for users to generate and share AI-generated video content.
Coupled with the roll-out of Sora 2, its new audio-visual software model, the platform aims to shift AI from a mere productivity tool, into a space for social connection.
The move positions the tech behemoth directly against market leader TitTok, which has long dominated the short-form video scene.
“Prior video models are overoptimistic, they will morph objects and deform reality to successfully execute upon a text prompt”, OpenAI wrote in a blog post.
“For example, if a basketball player misses a shot, the ball may spontaneously teleport to the hoop. In Sora 2, if a basketball player misses a shot, it will rebound off the backboard”.
While Tiktok has become synonymous with viral trends and algorithm-driven video content, Sora hopes to set itself apart by creating an entirely AI-generated environment.
This includes a ‘cameo’ feature, which will allow people to authenticate their likeness to then drop themselves into AI-generated scenes.
This will let them control how their image is used on the app, which has become a crucial measure in a space where non-consensual AI content is a growing concern.
The Sora app emulates the TikTok style ‘for you page’ feed, where uses can scroll through algorithmically organised clips.
Unlike the Chinese-owned app, however, which stands on a passive consumption mode, OpenAI promises to emphasise creative agency, allowing its users to actively generate, remix and share content.
Videos are limited to just ten seconds, a nod to the growing short-form content trend.
A ‘user focused’ social media platform
Sam Altman’s firm is rolling out the platform with a degree of caution.
Access to the app has only been launched on an ‘invite only’ basis on Apple phones the US and Canada, whilst Android availability is to be expected in due course.
Priority has been handed to heavy users of the original Sora model and ChatGPT pro customers, as OpenAI seeks to seed its new platform with highly engaged, early adapters.
The tech giant’s approach to user identity and content rights has been carefully considered, with users being able to grant friends’ permission to include their likeness in videos, with full control to approve or delete content.
While copyright is left largely in the hands of rights holders, OpenAI will apply digital labelling and watermarking to all Sora-generated video, in a pledge to maintain transparency within AI-generated content.
Sora represents OpenAI’s most significant foray into social networking yet, aligning with a broader industry craze; Meta recently launched ‘Vibes’, yet another short-form AI video feed, while Tiktok continues to expand its creative tools and UK presence.
OpenAI’s differentiator, it has claimed, lies in its AI-first approach. Instead of merely providing tools for editing or sharing, it makes content creation itself an AI-driven social experience.
Yet, in terms of adoption, TikTok’s success is built on habitual engagement, network effects, and the gigantic scale of its user base.
OpenAI will have to balance replicating that appeal while maintaining strict safety standards, it’s success hinging on the balance between creativity and regulation, as well as the platform’s ability to foster meaningful interaction without lapsing into the mindless content loops critics associate with its competitors.