OpenAI launches AI-powered video generator Sora

OpenAI unveiled Sora, an advanced Artificial Intelligence (AI) video generator. The new tool allows users to create highly realistic videos from text, images, and even existing videos. However, due to overwhelming demand, new account creation for Sora has been temporarily paused. However, the platform has not disclosed the exact number of users who signed up on the first day but confirmed a significant influx.

Sora employs a diffusion model, which generates videos by starting with a base video resembling static noise and progressively refining it. This process allows the software to transform descriptive text instructions into detailed, visually accurate videos. Similar to DALL-E’s recaptioning technique, Sora is trained on publicly available, proprietary, and partner-sourced data, enabling it to interpret user instructions with high fidelity.

Users can create videos in widescreen, vertical, or square formats, with resolutions up to 1080p and a maximum length of 20 seconds. Sora also excels at editing and enhancing existing videos, such as filling in missing parts or extending endings.

The launch highlights OpenAI’s continued push to innovate in AI-driven creative tools, offering new possibilities for content creators, marketers, and filmmakers. While new users are temporarily unable to join, OpenAI has encouraged prospective users to stay tuned for updates on account availability.

With its ability to generate lifelike videos from text, Sora is poised to revolutionise how we create and consume visual content.