Artificial intelligence video generation is no longer just a pursuit of the most impressive visuals. For those looking to use AI for creating short films, commercials, or even feature-length movies, questions are becoming more pressing: Will characters remain consistent, will shots unfold as intended, and can the video tell a coherent story from beginning to end?
The latest update to Kling AI, version 4.0, offers tools designed to address these challenges. The main focus is on improving motion handling, providing creative control, and supporting longer narratives. The model now supports generating up to 30 seconds of video in a single pass, as well as up to 15 multimodal references (images, videos, audio) and 10 keyframes. This allows creators to more precisely control characters, scenes, actions, and camera movements.
Traditionally, AI video relied primarily on text prompts. However, accurately describing complex camera movements, character actions, and scene transitions often proved difficult. Even if individual frames looked convincing, their combination could lead to inconsistencies in character appearance or choppy actions. Multimodal references and keyframes in Kling 4.0 serve precisely to overcome these limitations, allowing users to specify not only the visual 'what' but also the directorial 'how'.
Essentially, Kling 4.0 shifts the focus from simply creating 'pretty' frames to ensuring directorial control and narrative coherence. This doesn't mean AI is now independently shooting films – plot development, shot selection, and final editing still require human involvement. However, the new capabilities make AI tools more predictable and useful for professional video production, where not only an individual striking scene is important, but also how it fits into the overall story.