OpenAI unveils revolutionary text-to-video AI model
Revolutionizing digital media, OpenAI’s Sora translates written prompts into visually compelling and detailed minute-long videos
OpenAI, the company behind the well-known AI tools ChatGPT and DALL-E, has introduced a groundbreaking text-to-video model named Sora.
This new technology allows users to generate hyper-realistic videos up to a minute long using simple text prompts.
Sora is currently in its testing phase and has captivated social media with its ability to create compelling characters, vivid emotions, and detailed scenes.
Notable demonstrations include a woman walking through Tokyo, mammoths in snowy landscapes and even more whimsical scenes such as marine animals in a bicycle race atop the ocean.
Sam Altman, the CEO of OpenAI, highlighted Sora’s capabilities by sharing videos created from public suggestions, including a grandmother making gnocchi and two golden retrievers podcasting on a mountain. These displays sparked amazement online, with users lauding the videos as “out of this world” and a significant advancement in AI technology.
However, the company acknowledges certain limitations in Sora, such as challenges in simulating complex physics and continuity errors. For instance, a character taking a bite from a cookie might not leave a visible mark afterward.
With the upcoming U.S. presidential election and the potential for misuse, OpenAI has prioritized safety. The company is working with experts to address concerns like misinformation, hateful content, and bias. They are developing tools, including a detection classifier, to identify videos generated by Sora and prevent inappropriate use.
Despite these breakthroughs, Sora remains unavailable for public use. OpenAI continues to consult with artists, designers and filmmakers, gathering feedback to refine the model. The company is also navigating ethical considerations, especially following lawsuits over the use of copyrighted materials to train its other AI models.
Source: Newsroom