COMIC: Agentic Sketch Comedy Generation
🤖 AI Summary
This work proposes a fully automated method for generating high-quality, diverse short comedic skit videos. By constructing a multi-agent collaborative system that emulates real-world film production workflows, the approach enables role-specific task assignment, iterative refinement, and competitive evaluation. A key innovation is the integration of a large language model critic trained on YouTube audience preferences to automatically align with and assess humor effectiveness. Combining large language models, a multi-agent framework, and video generation techniques, the method achieves near-professional levels in creativity, diversity, and comedic quality, significantly outperforming existing approaches and establishing state-of-the-art performance in automated comedic video synthesis.
Application Category
📝 Abstract
We propose a fully automated AI system that produces short comedic videos similar to sketch shows such as Saturday Night Live. Starting with character references, the system employs a population of agents loosely based on real production studio roles, structured to optimize the quality and diversity of ideas and outputs through iterative competition, evaluation, and improvement. A key contribution is the introduction of LLM critics aligned with real viewer preferences through the analysis of a corpus of comedy videos on YouTube to automatically evaluate humor. Our experiments show that our framework produces results approaching the quality of professionally produced sketches while demonstrating state-of-the-art performance in video generation.
Problem
Research questions and friction points this paper is trying to address.
AI-generated comedy
sketch video generation
automated humor evaluation
LLM-based critics
creative AI agents
Innovation
Methods, ideas, or system contributions that make the work stand out.
agentic system
LLM critics
comedy generation
automated video production
humor evaluation
🔎 Similar Papers
No similar papers found.
