Announced on Monday, Sora allows users to create AI-generated video clips from written prompts, marking a major expansion of OpenAI’s generative AI tools.
Previously accessible only to select artists, filmmakers, and safety testers since its debut in February, Sora’s release signals a new frontier in AI-powered storytelling.
Sora generates video content based on text descriptions. For example, OpenAI’s website showcases a video created from the prompt “a wide, serene shot of a family of woolly mammoths in an open desert,” depicting mammoths walking through sand dunes. OpenAI described Sora as a tool to “explore new forms of creativity” and “push the boundaries of video storytelling.”
Known for ChatGPT, OpenAI has also developed image-generation tools like DALL-E and is exploring voice cloning. Backed by Microsoft, the company is valued at nearly US$160 billion and remains a leader in the burgeoning AI market.
While promising, Sora has faced criticism over technical flaws, such as unrealistic physics and visual defects. Concerns over potential misuse have also been raised, including disinformation and deepfakes. OpenAI says it has implemented safeguards, such as limiting uploads of real individuals and blocking harmful content like sexual deepfakes and child exploitation materials.
Sora remains unavailable in the UK and Europe due to compliance with copyright laws, including the Online Safety Act and GDPR. Additionally, OpenAI suspended access to the tool earlier this month after activists exploited a vulnerability to bypass restrictions. The group accused the company of exploiting artists and warned against the implications of AI-generated content on creative livelihoods.
Despite these hurdles, Sora is accessible to paid OpenAI subscribers in the US and other select countries, with the company anticipating high demand as users explore its capabilities.
