OpenAI unveils Sora: Text-to-video AI system

OpenAI teased a new artificial intelligence system that can create realistic-looking videos based on text prompts from users, making it the latest AI company to embrace generative video technology. 

The AI system, called Sora, can quickly create videos up to a minute long that represent “complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background,” the company said in a blog post Thursday. 

• Also read: 59% of Indian enterprises have actively deployed AI: Report

The tool will initially be made available to “a limited number of creators,” OpenAI Chief Executive Officer Sam Altman wrote in a post on X. OpenAI is also granting access to a team of experts tasked with assessing Sora’s safety before incorporating it into the company’s products. 

Meta Platforms Inc., Alphabet Inc.’s Google and Runway AI have previously built text-to-video generators. While this technology could speed up the creative process, it also could add to concerns about AI upending the livelihoods of artists and spreading misinformation, particularly in a pivotal election year. 

• Also read: USIBC launches task force to align Indian, American leadership on AI

As part of its preparations for rolling out the product, OpenAI said it’s building tools to help detect when a video was generated by Sora. But AI companies have previously struggled to create foolproof AI detection systems. Last year, OpenAI shelved a classifier that had been intended to detect AI-generated writing, citing its “low rate of accuracy.”

More stories like this are available on bloomberg.com