Runway says its latest AI video model can actually generate consistent scenes and people

AI startup Runway says its latest AI video exemplary tin make accordant scenes and radical crossed aggregate shots, according to an announcement. AI-generated videos tin conflict with maintaining accordant storytelling, but Runway claims connected X that the caller model, Gen-4, should let users much “continuity and control” portion telling stories. 

Currently rolling retired to paid and endeavor users, the caller Gen-4 video synthesis exemplary allows users to make characters and objects crossed shots utilizing a azygous notation image. Users indispensable past picture the creation they want, and the exemplary volition past make accordant outputs from aggregate angles.

As an example, the startup released a video of a pistillate maintaining her quality successful antithetic shots and contexts crossed a assortment of lighting conditions.

The merchandise comes little than a twelvemonth aft Runway announced its Gen-3 Alpha video generator. That exemplary extended the magnitude of videos users could produce, but also sparked controversy arsenic reportedly had been trained connected thousands of scraped YouTube videos and pirated films.