OpenAI’s “Sora” Model

 

OpenAI’s “Sora” Model

GS-3: Science & Technology

(UPSC/State PSC)

Important for preliminary exam:

ChatGPT, Generative Artificial Intelligence (GenAI) model “Sora”.

Important for Main Exam:

About Sora, Mechanism, Significance, Works of Sora model, Concerns & Limitations of OpenAI Sora, Future Outlook of Sora.

19 February 2024

Why in News:

Recently, OpenAI CEO Sam Altman, the creator of the revolutionary chatbot ChatGPT, has launched a new Generative Artificial Intelligence (GenAI) model “Sora”.

About Sora:

  • Sora means sky in Japanese, an imagery that evokes 'limitless creative potential', per the company's engineering team.
  • This new diffusion-based AI model is built on the foundation of the Transformer architecture, similar to larger language models like ChatGPT.
  • Sora is an OpenAI text-to-video model that generates videos with complex scenes that are up to a minute long.
  • Sora creates these videos from the user’s descriptive captions and still image prompts.
  • OpenAI has specified that the main goal of Sora is for real-world use; helping people solve problems that require real-world interaction by training AI to understand the physical world in motion.

Mechanism:

  • Diffusion models are used to generate high-quality images and videos in “Sora”.
  • Diffusion model is a physical process in which molecules move from high-concentration to low-concentration zones.

Significance:

  • Sora is an AI model that can create realistic and imaginative scenes from text instructions.
  • Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user’s prompt.
  • Sora is able to generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background.
  • It can create images and videos with near- accuracy on a given subject. It can create a video from an image and can also fill in gaps in existing video clips.

Works of Sora model:

  • Text to Video: The Sora model is capable of producing videos up to one minute in length, ensuring high visual quality and adherence as per the user’s instructions.
  • Generate complex scenes: Sora can generate intricate scenes featuring multiple characters, various types of motion, and precise details of both the subject and the background.
  • Dynamic Impressions: It can understand how objects function in reality, interpret prompts accurately. It can generate engaging characters that convey lively emotions.
  • Multishot Avatar: Sora can also produce multiple shots within a single generated video that accurately persist characters and visual style.
  • Note: Currently, the Sora model is not available in OpenAI’s products. It will be accessible after all safety checks are completed.

About Generative Artificial Intelligence (GenAI):

  • Generative AI uses Artificial Intelligence and Machine Learning algorithms to enable machines to generate new content (machine generated).
  • Systems use previously created content, such as text, audio, video, images, and code.
  • The term ‘Generative’ refers to the ability of the models to learn how to create new data rather than simply recognising it. For example, a generative model may learn how to generate images that resemble faces given a set of parameters (such as the eyes, hair, or skin colour etc.)

Concerns & Limitations of OpenAI Sora:

  • The current model has weaknesses. It may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect.
  • Like many AI applications, OpenAI Sora cannot be used to generate videos depicting violence, adult content, or that of real people or in the style of named artists. Upon its announcements, immediate questions have also been raised on copyright infringement by filmmakers and creative professionals, as well as the program’s safety and veracity – particularly in the spread of misinformation. Despite researcher Bill Peebles claiming that “training data is from the content we’ve (OpenAI Sora) licensed and also publicly available content”, several lawsuits regarding AI’s use of “publicly available” copyrighted content is currently underway.

Future Outlook of Sora:

  • Open AI is building tools to help detect misleading content such as a detection classifier that can tell when a video was generated by Sora.
  • Open AI will be engaging policy makers, educators and artists around the world to understand their concerns and to identify positive use cases for this new technology.

Source: The Hindu

----------------------------------

Mains Question

What is OpenAI Sora? Discuss significance, concerns & limitations of it.