Sora, OpenAI's video generation model, isn't available to the public yet primarily due to ongoing safety and ethical concerns. The model is capable of generating realistic videos from text prompts, which introduces the potential for misuse, such as creating deepfakes or misleading content. To address these risks, OpenAI is conducting extensive testing with "red teamers," who are experts tasked with identifying vulnerabilities related to misinformation, bias, and harmful content.

Additionally, OpenAI is working on developing safeguards like detection tools that can flag AI-generated videos and incorporating metadata to ensure transparency. These precautions are necessary to mitigate the potential for Sora to be used maliciously before it's made widely available​ 

 

The technology is being fine-tuned, and OpenAI may introduce it to the public later, likely with restrictions and guidelines similar to other AI models like DALL-E​