Google Research presents four frameworks for coherent long-form video
Google Research has introduced an AI video co-director designed to turn short generated clips into coherent, multi-shot stories lasting several minutes. The system addresses identity drift, such as changing clothing or scenery, and cascading failures in which an early faulty asset affects later scenes. It operates as a model-agnostic orchestration layer above Gemini and Veo, whose outputs retain SynthID watermarking.
The suite contains four frameworks. Co-Director uses bandit search to select creative and narrative settings, then combines storyboard, media-generation and judging agents. CANVAS maintains visual records of characters, locations and objects. A²RD generates videos segment by segment using multimodal memory, while VQQA creates visual questions and uses model critiques to revise prompts and select the strongest result.
Google reports that Co-Director scored 81.4 on GenAD-Bench and received a human rating of 3.96 out of 5. CANVAS improved background, character and prop consistency in its tests, and A²RD produced a continuous 10-minute film. Co-Director and A²RD code are publicly available; CANVAS is planned for release. The full pipeline is not a Google product.
Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.