OpenAI founder and CEO, Sam Altman announced a new product from the company. It is called Sora, an AI model that can create realistic and imaginative scenes from text instructions.
We were wondering when OpenAI was going to release its own video engine as many of its rivals already have one. It is here now and will be available to the public once it is safe for use. It is possible that OpenAI wanted to get things right before launching an AI video engine.
At the moment, the quality of the one-minute videos Sora AI output is putting pressure on competitors. You can check out Sora through their official page. According to the company, it is capable of producing realistic and imaginative scenes from single text prompts, like other similar products in the market.
They claim that their AI understands and can simulate the physical world in motion. This allows the model to create scenes featuring multiple people, objects, different types of movements, facial expressions, and textures with great detail.
This model is also multimodal. Once it is live, users will be able to use images for the base of a video. The content inside the image will be animated with attention to detail.
Sam interacted with users on X and created some videos by following the prompts given by X users. Here are some examples of videos shared by Sam.
The AI model is not perfect but they are getting there. OpenAI admits that it has weaknesses. The model can have a hard time simulating the physics of an object or get confused from left to right and so on.
Safety is an important part of making such products available in the market. OpenAI says, “We are working with red teamers — domain experts in areas like misinformation, hateful content, and bias — who will be adversarially testing the model.” They are also building tools to help detect misleading content as well.
Right now, there is no official word from the company about the launch.
