OpenAI, in collaboration with Microsoft, launched its first text-to-video tool on Thursday, named “OpenAI Sora”, similar to Google’s Lumiere. Sora’s availability is also limited like Google’s Lumiere. However, unlike Google’s Lumiere, which can generate videos up to 5 seconds long only, OpenAI’s Sora can generate videos up to 1 minute long.
Both OpenAi Sora and Google Lumiere are text-to-video Artificial Intelligence tools that have the power to transform the entire world of tech in a few years.
OpenAI Sora Vs Google Lumiere
According to OpenAI’s blog post,
“Sora is capable of creating complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background.”
OpenAI also claimed that the new AI model
“can understand how objects exist in the physical world, and accurately interpret props and generate compelling characters that express vibrant emotions.”
Cornell University described Google’s Lumiere as, “a text-to-video diffusion model designed for synthesizing videos that portray realistic, diverse, and coherent motion — a pivotal challenge in video synthesis. A Space-Time U-Net architecture that generates the entire temporal duration of the video at once, through a single pass in the model. This is in contrast to existing video models which synthesize distant keyframes followed by temporal super-resolution — an approach that inherently makes global temporal consistency difficult to achieve.”
OpenAI Sora vs Google Lumiere: Use and Availability
OpenAI has not yet launched Sora to the general public. Before launching, the company will start its outreach program with policymakers and security experts, where it will analyze the generation of information from the model so that any hateful content is not generated.
Google’s Lumiere is also undergoing its development. Thus, it is not available to the public yet. However, constant refining and testing of the model is under process by the researchers. They are operating with hopes that shortly, Lumiere will be available to a wider audience.
OpenAI Sora vs Google Lumiere: Shortcomings
OpenAI announced that the present model of Sora has certain shortcomings. Sora may fail to understand the specific law of physics behind simulating a complex scene. Moreover, it may not understand the linkage of cause and effect. The company gave an example,
“a person might take a bite out of a cookie, but afterward, the cookie may not have a bite mark.”
The Google Research team acknowledged certain risks that could be associated with Lumiere, as proposed by its software. The team mentioned that Lumiere possesses the risk of misuse of information for the creation of harmful or fake content. Thus the team will emphasize inculcating tools that would detect malicious content and biases to ensure safe and fair use. Moreover, people are also concerned about the source from where Lumiere collects its training data. The risk of Lumiere replacing human animators also exists.
Also Read: Napoleon vs Amazon: How To Join Amazon Class Action Lawsuit?
