OpenAI Unveils Sora, a Text-to-Video AI Model
OpenAI introduces Sora, a new generative AI model that creates realistic 60-second videos from simple text prompts. The tool still faces limitations with spatial details and cause-and-effect logic.
Artificial intelligence leader OpenAI introduces a new AI model called Sora that creates realistic and imaginative 60-second videos from quick text prompts. The company states that Sora generates scenes with multiple characters, specific types of motion, and detailed backgrounds while understanding how these elements exist in the physical world.
While text-to-video models already exist, Sora stands out because of the impressive length and accuracy of its generated clips, according to industry analysts. Experts believe this technology significantly impacts digital entertainment markets by enabling personalized content creation, such as generating short scenes to support television narratives.
Despite its advanced capabilities, OpenAI admits Sora remains a work in progress with clear weaknesses in spatial details and cause-and-effect logic, such as struggling with left and right orientation or failing to show a bite mark after a character eats a cookie. The company currently focuses on safety as it continues developing the model for real-world problem solving.