Sora: OpenAI unveils text-to-video AI tool
Access is currently limited to researchers and content curators
OpenAI, creator of ChatGPT, has unveiled Sora, a tool capable of crafting lifelike 60-second videos from simple text prompts.
By providing text prompts and photos, individuals can instruct Sora to produce videos tailored to their specifications.
According to OpenAI, the tool not only generates videos based on text prompts but also has the capability to use still images or existing footage as references, enhancing its flexibility and utility for content creators.
Illustrating the tool's capabilities, OpenAI showcased a sample prompt envisioning "a movie trailer featuring the adventures of the 30-year-old space man wearing a red wool knitted motorcycle helmet, blue sky, salt desert, cinematic style, shot on 35mm film, vivid colours."
Sora, named after the Japanese word for sky, quickly translated this text into a visually advanced depiction, underscoring AI's improving ability to comprehend and execute complex instructions, assuming the demos are accurate.
OpenAI CEO Sam Altman engaged with users on X (formerly Twitter), sharing videos generated by Sora in response to their prompts. The company watermarked Sora's outputs with C2PA tags to show they are AI-generated and to allow their provenance to be tracked.
"Sora has a deep understanding of language, enabling it to accurately interpret prompts and generate compelling characters that express vibrant emotions," OpenAI explained.
"The model understands not only what the user has asked for in the prompt, but also how those things exist in the physical world."
Access to Sora is currently limited to researchers and content curators, as a precautionary measure to ensure compliance with OpenAI's security policies.
Prohibiting content such as extreme violence and sexual imagery, the company claims it is taking measures to uphold ethical standards and prevent misuse of the technology.
"We'll be engaging policymakers, educators, and artists around the world to understand their concerns and to identify positive use cases for this new technology," OpenAI said.
OpenAI remains tight-lipped regarding the intricacies of Sora's workings, emphasising instead its focus on security and refinement before wider public release.
OpenAI's previous milestones include the introduction of DALL-E, a still image generator, and ChatGPT, the conversational AI platform, which saw widespread and rapid adoption.
Commenting on OpenAI's new AI tool, Dr Andrew Rogoyski, director of innovation and partnerships at the University of Surrey, said: "OpenAI's Sora system potentially marks a ChatGPT moment for video. In the same way that the text-based ChatGPT launched just over a year ago has transformed and disrupted text-based work, from scriptwriting to email marketing, the Sora system could do the same for video."
He commended OpenAI for recognising the potential for harm and restricting its use initially.
"The idea that an AI can create a hyper-realistic video of, say, a politician doing something untoward should ring alarm bells as we enter into the most election-heavy year in human history, with over 60 democratic elections in 2024 and half the planet's population voting."
However, Rogoyski wondered whether the UK's new AI Safety Institute, announced by Rishi Sunak last November at Bletchley Park, "and enjoying a commitment from the big AI firms to share their breakthroughs," had been made aware of Sora before its release.
This article originally appeared on our sister site Computing.