Producing video with artificial intelligence is nothing new. However, the biggest challenge encountered in using such models is to provide consistency among different scenes. It is very difficult to transfer a human character to different scenes with the same features. These objects are also taking place. However, Runway, one of the leading companies in this regard, says that with the newest model Gen-4, it has overcome such problems.

Runway is forcing the boundaries

This model, which is offered to the company’s paid individual and corporate customers, promises significant developments in character consistency, stage compliance and realistic movements compared to previous generations. According to Runway’s statement, Gen-4, visual references and text instructions can offer consistent style, subject and space. This allows users to create the same character in different angles and under lighting conditions. Users can produce realistic and dynamic videos by identifying stage compositions.

Runway is a little different from other artificial intelligence companies. While big competitors such as OpenAI and Google compete with each other, Runway is trying to make a difference with Hollywood studios and budgets for AI -supported film projects. The company points out that the model has reached a new level in simulating real world physics.

As in every artificial intelligence model, the data sets in which Gen-4 is trained is also a question of curiosity. Runway avoids explaining educational data in order not to lose its competitive advantage. In the world of artificial intelligence, this situation brings to mind the “royalties .. Runway defends himself within the framework of the “fair use” doctrine in the face of these concerns.