OpenAI, supported by Microsoft, has announced its development of Sora, a software capable of creating one-minute videos from text prompts. According to the company’s statement on Thursday, Sora is presently accessible for red teaming purposes, aiding in the identification of flaws within the AI system. Additionally, it’s available for utilization by visual artists, designers, and filmmakers, facilitating feedback on the model’s performance.
Sora, as stated in a recent blog post by the company, is not only capable of generating videos from text prompts but can also animate still images.
Following in the footsteps of OpenAI’s ChatGPT, which garnered attention upon its release in late 2022 for its proficiency in composing emails, writing code, and crafting poetry, Sora represents another step forward in the field of generative AI.
Meta Platforms, the parent company of Facebook, enhanced its image generation model, Emu, last year, introducing two AI-based features for editing and generating videos from text prompts. This move underscores Meta’s ambition to compete with industry giants like Microsoft, Google, and Amazon in the rapidly evolving landscape of generative AI.
While Sora is still a work-in-progress, the company acknowledges potential challenges such as the model’s occasional confusion with spatial details in prompts and difficulties in adhering to specific camera trajectories.
OpenAI has also revealed ongoing efforts to develop tools capable of discerning whether a video was generated by Sora, further illustrating the advancements being made in this innovative field.