Image - Sora,OpenAI
On Thursday, OpenAI, the creator of ChatGPT and DALL-E, introduced Sora, a text-to-video diffusion model, marking a significant leap in generative artificial intelligence capabilities.
"Sora" names after the Japanese term for "sky," has the capability to generate lifelike videos lasting up to a minute, tailored to user specifications in terms of content and aesthetic. As per a company announcement, Sora is also equipped to craft videos from single images or augment existing footage with fresh content.
Among the initial examples provided by the company, one video was created in response to the prompt: “Animated scene features a close-up of a short fluffy monster kneeling beside a melting red candle. the art style is 3d and realistic, with a focus on lighting and texture. the mood of the painting is one of wonder and curiosity, as the monster gazes at the flame with wide eyes and open mouth. its pose and expression convey a sense of innocence and playfulness, as if it is exploring the world around it for the first time. the use of warm colors and dramatic lighting further enhances the cozy atmosphere of the image.”
Sora is now accessible to red teamers, who are experts tasked with adversarially testing the model for potential harms and risks. Additionally, a select group comprising visual artists, designers, and filmmakers has been granted access to provide feedback aimed at enhancing the model's utility for creative professionals.
Since the launch of ChatGPT in November 2022, OpenAI has been rapidly developing generative AI tools. This journey has seen the release of GPT-4, voice and image prompts, and the latest DALL-E 3 image model, all seamlessly integrated within ChatGPT. Furthermore, OpenAI's API has significantly influenced the AI industry by empowering companies and developers to create their own generative AI tools.
With the introduction of Sora, OpenAI is poised to elevate AI capabilities to a new level with video generation.
While other video-generating models exist, none match Sora's purported ability to produce realistic and intricate videos. Meta offers a tool for generating short video clips, and Google is actively researching its own text-to-video model, albeit still in the experimental phase.
Sora enables users to create videos up to one minute long, featuring detailed scenes and multiple characters. The announcement showcases clips demonstrating Sora's capabilities, including footage of an SUV navigating a winding mountain road and "historical" scenes depicting California during the gold rush era.
Addressing safety concerns, OpenAI emphasizes its commitment to implementing measures to ensure responsible usage of Sora-generated content. These measures include labeling Sora-created videos in accordance with C2PA guidelines and employing existing safety protocols, similar to those applied to DALL-E, to filter out inappropriate or harmful text prompts.
Furthermore, OpenAI intends to engage policymakers, educators, and artists worldwide to understand their concerns and identify positive use cases for this groundbreaking technology. The company underscores the importance of real-world feedback in refining and enhancing the safety of AI systems over time.
In unveiling Sora, OpenAI is poised to revolutionize video content creation while prioritizing safety and responsible usage.